University of Southern California Dissertations and Theses

Human-environmental interaction: potential use of pupil size for office lighting controls

(USC Thesis Other)

Human-environmental interaction: potential use of pupil size for office lighting controls

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content HUMAN-ENVIRONMENTAL INTERACTION:
POTENTIAL USE OF PUPIL SIZE FOR OFFICE LIGHTING CONTROLS

By

Rui Zhu

______________________________________________________________________________

A Thesis Presented to the
FACULTY OF THE USC SCHOOL OF ARCHITECTURE
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF BUILDING SCIENCE

August 2014

Copyright 2014 Rui Zhu

2
Dedication
This thesis is dedicated unconditionally to my family. The thesis can not be realized without
their insistently support. My dad didn’t have a chance to see the completion of this thesis, but
without him, I would not receive the opportunity to receive great education and start this thesis.
My mom, sacrificed even more, bearing the sorrow and loneliness, to support me to finish all the
thesis work.
Thank you dad, thank you mom! I love you forever.

3
Acknowledgement
I would like first and foremost to acknowledge and thank Professor Joon-ho Choi for his
entirely involvement in this project. He opened my mind to this interesting research and guided
me in the very beginning to teach basic ideas and train key technical skills. He offered great help
in shaping the structure of this research and ensuring the direction in the right way. He also
provided great support in financing the devices of the project, which made experiment possible.
Dr. Choi, you are not only an instructor, but also a friend of mine. Your guidance will always
benefit my life.
I would also like to acknowledge Professor Douglas Noble for his endless support to this
research. Professor Noble could always offer a positive altitude whenever there was an obstacle
in the way and encouraged me to fight against it. I really appreciated his participation in the
experiment and his valuable advice on the improvements of the project. Professor Noble is like a
grandpa of mine, sometimes restricted, but cared about me.
In addition, I would like to thank Professor Karen Kensek. Karen can always provide very
useful and interesting ideas to the project and is really careful about all the content of the thesis. I
would also like to appreciate Karen and her husband for the technical advice to the project.
Without them, I would suffer more difficulties.
Last but not least, I would like to thank all MBS family and my friends for supporting my
research project. Without you consistent help, I would never get enough data and finish this
thesis. I would also thank Master of Building Science Program for giving me the opportunity of
enjoying such wonderful time and completing this thesis project.

4
Table of Contents
Dedication ...................................................................................................................................... 2

Acknowledgement ......................................................................................................................... 3

List of Figures ................................................................................................................................ 6

List of Tables ............................................................................................................................... 11

Abstract ........................................................................................................................................ 13

Hypothesis .................................................................................................................................... 14

Chapter 1: Introduction of Study .............................................................................................. 15

1.1 - Problem ............................................................................................................................. 15

1.2 – Physiological Response .................................................................................................... 18

1.3 – Pupil Size .......................................................................................................................... 19

1.4 – Objective ........................................................................................................................... 19

Chapter 2: Background Research ............................................................................................. 21

2.1 –Human Health and Productivity Corresponding to Lighting Environment ....................... 21

2.2 – Achieving Better Lighting Environment and Design Methods ........................................ 22

2.3 – Human Pupil Sizes and the Potential Use ........................................................................ 25

2.4 – Conclusions for Background Research ............................................................................ 27

Chapter 3: Methodologies .......................................................................................................... 29

3.1 - Scope of Work ................................................................................................................... 29

3.1.1 – Lighting Parameters ................................................................................................... 31

3.1.2 – Task Types ................................................................................................................. 33

3.1.3 – Questionnaire and Performance Test ......................................................................... 33

3.1.4 – Pupil Size Parametric Data ........................................................................................ 36

3.2 - Experimental Chamber Setup ........................................................................................... 37

3.2.1 – Chamber Design ........................................................................................................ 37

3.2.2 – Lighting Fixtures ....................................................................................................... 38

3.3 – Research Tools and Sensor Devices ................................................................................. 39

3.3.1 – Sensory Devices: Illuminance Meter, Luminance Meter and HDR camera ............. 39

3.3.2 –DAQ device ................................................................................................................ 42

3.3.3 –ASL Mobile Eye XG .................................................................................................. 43

3.4 - Adopted Software .............................................................................................................. 44

3.4.1 – LabVIEW ................................................................................................................... 44

3.4.2 – Programming Logic ................................................................................................... 45

3.4.3 – Photolux 2.1 ............................................................................................................... 46

5
3.4.4 – Minitab ....................................................................................................................... 48

3.5 - IRB Preparation ................................................................................................................. 49

3.6 – Preliminary Study and Results ......................................................................................... 50

3.7 – Experiment Rounds .......................................................................................................... 52

3.8 – DATA Analysis ................................................................................................................ 54

Chapter 4: Study Results ........................................................................................................... 56

4.1 – Pilot Study Results ........................................................................................................... 56

4.2 – First Round Results: Low light color temperature condition, Computer task type .......... 61

4.3 – Second Round Results: High light color temperature condition, Computer task type ..... 66

4.4 – Third Round Results: High light color temperature condition, Paper task type .............. 70

Chapter 5: Data Analysis and Discussion ................................................................................. 74

5.1 – First Round ....................................................................................................................... 74

5.2 – Second Round ................................................................................................................... 91

5.3 – Third Round .................................................................................................................... 104

5.4 – Discussions between different rounds ............................................................................ 118

5.4.1- Checking Consistency of Previous Observations ...................................................... 118

5.4.2- Findings Observed from Comparisons between Experiment Rounds ....................... 120

5.5 – Summary ......................................................................................................................... 121

Chapter 6: Conclusions of Study ............................................................................................. 124

6.1 – Illuminance, Sensations and Pupil Size .......................................................................... 124

6.2 – Further Conclusions based on Physiological Features of Human Subjects ................... 125

6.3 – Color Temperature and Task Type ................................................................................. 126

6.4 – Potential Use of Findings ............................................................................................... 127

Chapter 7: Future Work .......................................................................................................... 129

7.1 – Possible Improvements on Participants .......................................................................... 129

7.2 – Possible Improvements on Lighting Parameters ............................................................ 129

7.3 – Possible Improvements on Hardware and Software ....................................................... 130

7.4 – Strategy Development .................................................................................................... 132

Bibliograhpy ................................................................................................. 错误! 未定义书签。

6
List of Figures
Figure 1.1 View in a Lighting Simulation Software: Autodesk 3DS MAX ................................. 17

Figure 1.2 Normal Image vs. HDR Image .................................................................................... 18

Figure 3.1 Conceptual diagram of the research methodologies .................................................... 29

Figure 3.2 Sample Section of Designed Questionnaire ................................................................ 35

Figure 3.3 Participant Wearing ASL Mobile Eye-XG Device ..................................................... 36

Figure 3.4 Diagrammatic plan of Modifying Chamber ................................................................ 38

Figure 3.5 Dimmable LED Lamp ................................................................................................. 39

Figure 3.6 OMEGA HHLM-1 ...................................................................................................... 42

Figure 3.7 Coolpix 8400 ............................................................................................................... 42

Figure 3.8 Cooke cal-SPOT 401 ................................................................................................... 42

Figure 3.9 NI USB-6008 ............................................................................................................... 42

Figure 3.10 ASL Mobile Eye XG ................................................................................................. 43

Figure 3.11 Fisheye View of the Chamber ................................................................................... 44

Figure 3.12 Front Panel of Designed Program in LavVIEW ........................................................ 46

Figure 3.13 Block Diagram of Designed Program in LabVIEW .................................................. 46

Figure 3.14 Four images taken at different exposure settings for same illuminance setting
displayed in Photolux before combining. ..................................................................................... 47

Figure 3.15 Processed and analyzed image after combining four images taken at different
exposure settings for same illuminance setting in Photolux. ........................................................ 48

Figure 3.16 Sample interface of Minitab ...................................................................................... 49

Figure 3.17 Human Research Curriculum Report ........................................................................ 50

Figure 3.18 Images taken for some participants in the experiment .............................................. 54

Figure 3.19 Data Analysis Process ............................................................................................... 55

Figure 4.1 Illuminance range per visual sensation of each individual (Pilot study). .................... 57

Figure 4.2 Overall illuminance distribution per visual sensation of all individuals (Pilot study). 58

Figure 4.3 Ranges of standardized pupil size per visual sensation of each individual (Pilot study).
....................................................................................................................................................... 59

7
Figure 4.4 Overall standardized pupil size distribution per visual sensation of all individuals
(Pilot study). .................................................................................................................................. 60

Figure 4.5 Illuminance range per visual sensation of each individual (First round). ................... 63

Figure 4.6 Overall illuminance distribution per visual sensation of all individuals (First round). 64

Figure 4.7 Ranges of original pupil size per visual sensation of each individual (First round). .. 65

Figure 4.8 Overall original pupil size distribution per visual sensation of all individuals (First
round). ........................................................................................................................................... 66

Figure 4.9 Illuminance range per visual sensation of each individual (Second round). ............... 68

Figure 4.10 Overall illuminance distribution per visual sensation of all individuals (Second
round). ........................................................................................................................................... 69

Figure 4.11 Ranges of original pupil size per visual sensation of each individual (Second round).
....................................................................................................................................................... 69

Figure 4.12 Overall original pupil size distribution per visual sensation of all individuals (Second
round). ........................................................................................................................................... 70

Figure 4.13 Illuminance range per visual sensation of each individual (Third round). ................ 72

Figure 4.14 Overall illuminance distribution per visual sensation of all individuals (Third round).
....................................................................................................................................................... 72

Figure 4.15 Ranges of original pupil size per visual sensation of each individual (Third round). 73

Figure 4.16 Overall original pupil size distribution per visual sensation of all individuals (Third
round). ........................................................................................................................................... 73

Figure 5.1 Standardized pupil size distribution in each subject’s test (“No.” indicates a subject ID)
(First round). ................................................................................................................................. 75

Figure 5.2 Overall standardized pupil size distribution per visual sensation to illuminance
intensity (first round). ................................................................................................................... 76

Figure 5.3 Interval plot of standardized pupil size per visual sensation to illuminance intensity
(First round). ................................................................................................................................. 77

Figure 5.4 Boxplot of standardized pupil size per visual sensation to illuminance intensity (First
round). ........................................................................................................................................... 78

Figure 5.5 Interval plot of comparisons of overall standardized pupil size per visual sensation
between eye color groups (first round). ........................................................................................ 80

8
Figure 5.6 Boxplot of comparisons of overall standardized pupil size per visual sensation
between eye color groups (first round). ........................................................................................ 81

Figure 5.7 Interval plot of comparisons of overall standardized pupil size per visual sensation
between age groups (first round). ................................................................................................. 83

Figure 5.8 Boxplot of comparisons of overall standardized pupil size per visual sensation
between age groups (first round) .................................................................................................. 84

Figure 5.9 Interval plot of comparisons of overall standardized pupil size per visual sensation
between myopic groups (first round). ........................................................................................... 86

Figure 5.10 Boxplot of comparisons of overall standardized pupil size per visual sensation
between myopic groups (First round) ........................................................................................... 87

Figure 5.11 Interval plot of comparisons of overall standardized pupil size per visual sensation
between gender groups (first round). ............................................................................................ 89

Figure 5.12 Boxplot of comparisons of overall standardized pupil size per visual sensation
between gender groups (first round) ............................................................................................. 90

Figure 5.13 Standardized pupil size distribution in each subject’s test (“No.” indicates a subject
ID) (second round). ....................................................................................................................... 92

Figure 5.14 Overall standardized pupil size distribution per visual sensation to illuminance
intensity (second round). ............................................................................................................... 93

Figure 5.15 Interval plot of standardized pupil size per visual sensation to illuminance intensity
(second round). ............................................................................................................................. 94

Figure 5.16 Boxplot of standardized pupil size per visual sensation to illuminance intensity
(second round). ............................................................................................................................. 95

Figure 5.17 Interval plot of comparisons of overall standardized pupil size per visual sensation
between eye color groups (second round). ................................................................................... 96

Figure 5.18 Boxplot of comparisons of overall standardized pupil size per visual sensation
between eye color groups (second round). ................................................................................... 97

Figure 5.19 Interval plot of comparisons of overall standardized pupil size per visual sensation
between age groups (second round). ............................................................................................. 98

Figure 5.20 Boxplot of comparisons of overall standardized pupil size per visual sensation
between age groups (second round) .............................................................................................. 99

9
Figure 5.21 Interval plot of comparisons of overall standardized pupil size per visual sensation
between myopic groups (second round). .................................................................................... 100

Figure 5.22 Boxplot of comparisons of overall standardized pupil size per visual sensation
between myopic groups (second round) ..................................................................................... 101

Figure 5.23 Interval plot of comparisons of overall standardized pupil size per visual sensation
between gender groups (second round). ..................................................................................... 102

Figure 5.24 Boxplot of comparisons of overall standardized pupil size per visual sensation
between gender groups (second round) ...................................................................................... 103

Figure 5.25 Standardized pupil size distribution in each subject’s test (“No.” indicates a subject
ID). .............................................................................................................................................. 105

Figure 5.26 Overall standardized pupil size distribution per visual sensation to illuminance
intensity (third round). ................................................................................................................ 106

Figure 5.27 Interval plot of standardized pupil size per visual sensation to illuminance intensity
(third round). ............................................................................................................................... 107

Figure 5.28 Boxplot of standardized pupil size per visual sensation to illuminance intensity (third
round). ......................................................................................................................................... 108

Figure 5.29 Interval plot of comparisons of overall standardized pupil size per visual sensation
between eye color groups (third round). ..................................................................................... 110

Figure 5.30 Boxplot of comparisons of overall standardized pupil size per visual sensation
between eye color groups (third round). ..................................................................................... 110

Figure 5.31 Interval plot of comparisons of overall standardized pupil size per visual sensation
between age groups (third round). .............................................................................................. 112

Figure 5.32 Boxplot of comparisons of overall standardized pupil size per visual sensation
between age groups (third round) ............................................................................................... 112

Figure 5.33 Interval plot of comparisons of overall standardized pupil size per visual sensation
between myopic groups (third round). ........................................................................................ 114

Figure 5.34 Boxplot of comparisons of overall standardized pupil size per visual sensation
between myopic groups (third round) ......................................................................................... 115

Figure 5.35 Interval plot of comparisons of overall standardized pupil size per visual sensation
between gender groups (third round). ......................................................................................... 116

10
Figure 5.36 Boxplot of comparisons of overall standardized pupil size per visual sensation
between gender groups (third round) .......................................................................................... 117

Figure 5.37 Interval plot of comparisons of overall standardized pupil size per visual sensation
between color temperatures. ....................................................................................................... 120

Figure 5.38 Interval plot of comparisons of overall standardized pupil size per visual sensation
between task types. ..................................................................................................................... 121

Figure 5.39 Summary of average pupil size change in each category of first round. ................. 122

Figure 5.40 Summary of average pupil size change in each category of second round. ............ 122

Figure 5.41 Summary of average pupil size change in each category of third round. ................ 123

Figure 6.1 Conceptual strategy for automatic lighting controlling ............................................. 127

11
List of Tables
Table 3.1 Settings for lighting parameters .................................................................................... 32

Table 3.2 Settings for COOLPIX 8400 in “M” mode .................................................................. 40

Table 3.3 Aperture and Exposure Time Settings .......................................................................... 41

Table 3.4 Different Experiment Rounds ....................................................................................... 53

Table 4.1 Results of ANOVA with standardized pupil size data (left) and stepwise regression
based on the data of standardized pupil size, actual pupil size and illuminance. ......................... 61

Table 4.2 Demographic information of human subjects (First round) ......................................... 62

Table 4.3 Demographic information of human subjects (First round) ......................................... 67

Table 4.4 Demographic information of human subjects (Third round) ........................................ 71

Table 5.1 One-way ANOVA test: Standardized Pupil Size versus Sensation .............................. 79

Table 5.2 One-way ANOVA test: Standardized Pupil Size versus Sensation between Eye Colors
....................................................................................................................................................... 82

Table 5.3 One-way ANOVA test: Standardized Pupil Size versus Sensation between Age Groups
....................................................................................................................................................... 85

Table 5.4 One-way ANOVA test: Standardized Pupil Size versus Sensation between Myopic
Groups ........................................................................................................................................... 88

Table 5.5 One-way ANOVA test: Standardized Pupil Size versus Sensation between Gender
Groups ........................................................................................................................................... 90

Table 5.6 One-way ANOVA test: Standardized Pupil Size versus Sensation .............................. 95

Table 5.7 One-way ANOVA test: Standardized Pupil Size versus Sensation between Eye Colors
....................................................................................................................................................... 97

Table 5.8 One-way ANOVA test: Standardized Pupil Size versus Sensation between Age Groups
....................................................................................................................................................... 99

Table 5.9 One-way ANOVA test: Standardized Pupil Size versus Sensation between Myopic
Groups ......................................................................................................................................... 101

Table 5.10 One-way ANOVA test: Standardized Pupil Size versus Sensation between Gender
Groups ......................................................................................................................................... 103

12
Table 5.11 One-way ANOVA test: Standardized Pupil Size versus Sensation .......................... 108

Table 5.12 One-way ANOVA test: Standardized Pupil Size versus Sensation between Eye
Colors .......................................................................................................................................... 110

Table 5.13 One-way ANOVA test: Standardized Pupil Size versus Sensation between Age
Groups ......................................................................................................................................... 112

Table 5.14 One-way ANOVA test: Standardized Pupil Size versus Sensation between Myopic
Groups ......................................................................................................................................... 115

Table 5.15 One-way ANOVA test: Standardized Pupil Size versus Sensation between Gender
Groups ......................................................................................................................................... 117

13
Abstract
The goal of this research is to establish a visual environment diagnostic model based on the
occupant’s physiological responses for detecting improper ambient lighting conditions, a major
contributing factor to visual stress and work productivity in office workplace environments. The
human body, as a biological mechanism, naturally minimizes the effects of ambient
environmental stressors using its physiological autonomous nerve system. This system enables a
human’s pupils to dilate and contract, depending on visual sensations affected by the ambient
lighting conditions. An extensive experiment using human subjects will be conducted in an
environmental chamber on the University of Southern California campus. All parametric data
including human pupil sizes and lighting parameters will be categorized by age and ethnic origin,
to investigate and determine the most common features of pupil sizes per visual sensation among
individuals. Lighting parameters, including illuminance (lux), luminance (cd/m2), and
lighting-color-temperature (K), will be controlled and maintained for each volunteer subject
based on his/her task-type (computer-based or paper-based), which is most typical in the current
office environment.
This study will provide unique knowledge concerning how an occupant via his/her
physiological signal, i.e. pupil size can interact with the visual (lighting) environment. The
research outcome will be potentially applicable in reality to diagnose the lighting quality in
workplace environments, and to integrate an occupant’s pupil size information for the visual
environmental controls.

14
Hypothesis

Lighting design in the office building is always a crucial part for the whole indoor environment
quality. At present, in the US, most office buildings have adopted guidelines that were
empirically developed, primarily by the IESNA (Illuminating Engineer Society of North
America). But these guidelines that were empirically developed, mainly based on a conventional
paper-based bask-dominant environment. However, a computer-based work has become the most
popular task in the office since personal computer’s prosperity existed this 20 years. In this case,
current guidelines are not fit for the new working task. Furthermore, lighting simulation
programs and High-Dynamic Range (HDR) have also been used for detailed investigations such
as lighting design and glare analysis. However, human physiological features are not considered
in any of these current approaches. Regarding the biological function, pupil size could be
potentially used in the research to establish a visual quality assessment tool based on measuring
an individual’s pupil size so as to ensure one aspect of visual comfort in the built environment.

15
Chapter 1: Introduction of Study
Buildings have consumed almost 40% of all the energy consumption in the United States. The
number seems to be larger when it talks about the whole world. However, among all types of
buildings, the commercial or office buildings occupy large ratio of energy consumption. Lighting,
as one of the most important key components in the indoor environmental quality, nearly the
most significant one for the office environment, has a great relationship in energy consumption
as well as huge effect on human health and productivity.
Designing good lighting conditions attracts favorable attention from both architect and
engineering. Architects would have artistic lighting for a better presentation of the design:
highlighting details, dividing zones, creating a unique atmosphere and so on. While, for
engineers, their job is to ensure there is at least enough lighting for occupants to accomplish their
work. The occupants do not need to pay extra focus on getting right position for better lighting,
or feel tired due to improper lighting environment, which could be either too dark or too bright
while causing problems like glare. Demands from architects and engineers need realistic
solutions. Design guidelines or strategies are most appreciated especially when they could ensure
basic lighting level for high productivity as well as strengthen architecture artistic highlights.
Maintaining most ideal lighting environment is very impressed these days in the industry and
academia.
1.1 - Problem
Although great effort has been made for better indoor environment quality, most indoor
environmental components are managed based on pre-defined human comfort formulas and not
on the actual building occupants’ needs. Typically, environmental formulas are attained by
calculations and adopted to system industry standards and guidelines. As a result, an
16
individual‘s comfort is easily affected negatively, and unsatisfactory ambient conditions end up
with affecting the occupants’ work productivity and environmental health. Lighting, of all of
these factors, is most significantly related to the occupants’ visual comfort, which is instantly
affected and is easily vulnerable due to its immediate sensitivity. Recent studies have reported
that 65% of building occupants express that their workplace lighting conditions as inappropriate.
These occupants also report that they have considerable glare problems in their workplaces,
which can lead to serious visual stress(Irlen 1991). Despite its significance, most office buildings
have adopted empirically developed guidelines established mainly by the Illuminating
Engineering Society of North America (IESNA). Overall, these guidelines were designed mainly
based on general paper-based task environments which suggest that lighting guidelines may not
satisfy each individual lighting preference, and may result in unnecessary glare. Current
technical tools, such as lighting simulation and photo-based analysis (i.e. High Dynamic Range
image) have no functional feature to estimate a user's visual sensations in real time.
There are many lighting simulation software used in Building Information Modeling (BIM)
for daylight, including Ecotect, Radiance, Daysim and 3ds Max. Ecotect is mostly used for early
daylighting design, for the second stage, Radiance could provide more accurate daylighting
analysis; Furthermore, Daysim offer functions to control daylighting with presenting
performance at the same time. While, 3ds Max will create outstanding daylighting visualization
so as to explain lighting condition more directly. For artificial lights, there is less choice,
however, 3ds Max and AGI are considered as a good one.(“Lighting Analysis in BIM |
Sustainability Workshop” 2014) Figure 1.1 shows a simulation interface in the software.
17

Figure 1.1 View in a Lighting Simulation Software: Autodesk 3DS MAX
(“Light+Architecture: Lighting Simulation Software: Autodesk 3DS MAX” 2011)
Another popular tool in lighting design is High Dynamic Range (HDR) image. HDR could
achieve a higher dynamic range of luminosity and represent more accurately than commonly
used digital images by mostly merging multiple low dynamic range (LDR) or standard dynamic
range (SDR) photographs. HDR is good because of containing huge amount of information.
However, it takes too much time and effort to collect enough images, which are also not cost
efficient for public use. Figure 2 presents a comparison between HDR image and normal image.
18

Figure 1.2 Normal Image vs. HDR Image
(“6 Tips for Taking Better Macro Photos with the iPhone Camera” 2012)
What’s worse, neither of the two major tools can obtain real time function to estimate
occupants’ visual sensation. That is another reason figuring out a new way for aiding lighting
design and control is highly demanded.
1.2 – Physiological Response

The human body has an autonomic function that regulates its physical responses to minimize any
environmental stress, such as hot or cold temperatures, or excessively bright. For example,
depending on the intensities of various stressors, the skin on a human body could sweat or
control the surface body temperature to balance heat losses or gains caused by ambient thermal
conditions, and pupil sizes could shrink or dilate in response to variations in light. Therefore, this
research adopted human pupil sizes as a feasible physiological signal to estimate visual sensation
conditions (based upon the principle of reverse engineering) that could illustrate subjective
lighting sensations as a function of objectively measured physiological signals. The result would
19
be a novel method for visual quality assessment, such as a lighting simulation program and
high-dynamic images, as compared with conventional methods that have primarily depended on
pre-assumed human environmental reactions, instead of real human physiological responses.
1.3 – Pupil Size
Located in the central iris of human eye, the pupil is a hole which allows light to come into the
retina (Cassin & Solomon, 1990). It always shows black as the light coming into the pupil are
absorbed by the tissues or absorbed within the eye after diffuse reflections. The anatomical pupil
serves as aperture and iris as the aperture stop. The iris is consisted mainly of smooth muscle,
surrounding the pupil. Iris controls the amount of light entering the pupil by changing its size.
The pupil gets narrower in the light but wider in the dark. The diameter gets to 3 to 5 mm
when exposed to bright and respectively, to the maximum of 4 to 9 mm. Age is believed to have
significant effect on maximal pupil size. For instance, the diameter of the pupil could be 4mm to
9 mm when in a dark environment at age of 15. However, the average pupil size decrease at a
non-steady rate when human is older than 25.(“Event Horizon Volume 3 6 Aging Eyes and Pupil
Size” 2014; Winn et al. 1994) More information would be discussed more to indicate the
importance and possibility of using pupil size for the lighting control and improving human
health.

1.4 – Objective
The first purpose of this research is to establish a relationship between lighting conditions and
human pupil size, especially, the correlation between the illuminance levels and pupil size
change. By understanding effect of lighting condition on the pupil size, the possibility of
20
adopting pupil size serving as an indicator for the use of automatic lighting control could be
discussed.
The second objective of this study is to understand the variation of human pupil sizes among
different physiologically categorized people and to investigate the difference between raw pupil
sizes and normalized data. Human subjects differentiate from each other due to physiological
features. So does pupil. Pupil of people could behave significantly differently from each other
under the same lighting condition which indicates a various demand for the lighting. Therefore,
to check the variation of human pupil sizes is also very important to the study.
The third aim is to establish a visual quality assessment tool based on reading an individual
pupil size in order to demonstrate the potential use of pupil size for assuring visual comfort in the
office environment. Visual quality assessment depends seriously on huge amount of pupil size
data in different lighting environments and also a great variety of human subjects. With the
established visual quality assessment tool, it will be easier to judge satisfactions or comfort
feeling level of set lighting condition. As there are different requirements from individuals, an
optimal control strategy for office lighting setting would be decided based on e basis of all
occupants in that space. These strategies will help maintain a comfort lighting condition.
Meanwhile, practicing all related devices and software used in the project could be another
objective. Learning graphic programming in LabVIEW would benefit author a lot in future
engineering work. Furthermore, physical installation and related lighting device and of course
Mobile Eye XG would consolidate strong foundation of research skills related to lighting.

21
Chapter 2: Background Research
This chapter reviews previous studies and other sources relevant to lighting parameters
discussion and the use of pupil size in the lighting design area. Reviewing the potential benefits
and problems associated with pupil size and other important factors will help determine the
scope of the work in the study. Background information pertinent to the main areas of this thesis
was studied to help understand lighting parameters and function of pupil size as well as its
potential use in lighting design and control.
2.1 –Human Health and Productivity Corresponding to Lighting Environment
In order to understand better about the importance of proper-designed lighting in the office
environment, many studies have been conducted to investigate the correlations between human
health and productivity and different lighting conditions. A study conducted by Cornell
University (Hedge, Sims Jr., and Becker 1990) reported some findings of the offices which use
computers regarding the relationship of productivity, satisfaction and visual health of employees
and office lighting conditions in the background research. Complaints about light were collected
from 68% of workers in their offices based on the information provided by The American
Society of Interior Designers. A better lighting was demanded by 79% of VDT (Video Display
Terminal) users according to a study of Silicon Valley. A Louis Harris study in 1989 pointed out
eyestrain ranked as number one health hazard in the office which is ahead of radiation and
asbestos. All dissatisfaction discussed above is difficult to ignore which indicates potential
lighting problems. Better office lighting should be worked out to maintain the visual comfort for
the human subjects.
Not only in the surveys dissatisfaction was given to the lighting, but also other researchers
explained how lighting influence productivity. “We know that lighting affects people
22
psychologically and physiologically.”(Dilouie 2003) Visual impression occupies 80 to 85
percent of the entire process of learning about the world. However, perception depends on
lighting that makes it possible to for the visual purpose. Since people spend a large portion of
their time indoor, lighting is in charge of human being’s predominant perception of the world. It
has been long claimed that some lighting design methods are better as the lighting quality can
improve employees’ satisfaction. Nowadays, it is much harder to evaluate worker productivity in
offices, making satisfaction more important than before serving as a metric. Regarding the
important role in assessing performance and productivity of participant in the experiment, a
satisfaction survey was given to each participant in the experiment during the whole project.
2.2 – Achieving Better Lighting Environment and Design Methods
Many researches have been done to figure out better lighting methods in the office including
changing the lamp type, workplace layout, color temperature, illuminance and many other
potential elements in lighting control.
In the Cornell Study (Hedge, Sims Jr., and Becker 1990) as mentioned above, it was
designed to decide if applying a lensed indirect uplighitng system or a parabolic downlighting
would create any difference in the “visual comfort, satisfaction, health or productivity of
computer workers”. The research reported twice frequent complaints of tired eyes and
concentration problems in the indirect uplighting group than the parabolic group. Furthermore, it
was more bothersome in the parabolic than in the lensed indirect, and visual discomfort problems
under parabolic lighting conditions cut into worker productivity.
And in the research set up by Craig Dilouie, the effect of different forms of realistic office
lighting on the performance and health of employees in the offices was studied (Dilouie 2003).
23
Variables in the study included personal control for lighting, room surface brightness and so on.
The study concluded that direct/indirect fixtures was reported more comfortable than lensed and
parabolic troffers; People reported better lighting quality, satisfaction and showed higher
performance of attention and productivity when there was a dimmer for them to control; People
showed positive attitude to the job and working environment when they felt more satisfied with
lighting quality; Better task ability would be achieved by improving visibility in the lighting and
task conditions.
In summary, the study conducted by Dilouie found that people would feel more satisfied
when it was brighter on non-task room surfaces and private dimming control was available to set
preferred light levels for the occupants themselves. Satisfaction with lighting rate would result in
more focused attention, positive attitude and higher productivity during the work. Results also
indicated that more than 25% were not satisfied with standard lighting. It also concluded that by
combining direct/indirect lighting with private dimmer and perimeter wallwashing, lighting was
reported most comfortable and greatest satisfaction and motivation in a large amount of
population, that served as a great option for the owner to achieve economic benefits as well as
healthy and productivity.
Besides the type of lighting and the availability of a personal control, other lighting
parameters studies drew a great attention in the lighting research field, and among those
researches, the color spectrum or color temperature has been put the most attention as it is
believed to have the closest correlation with the human visual comfort and productivity.
A report done by the Pennsylvania Power and Lighting Company (PP&L) pointed out that
there was a high error rate in the N3 Drafting Department which was caused by employees were
24
not doing to their best because of the inadequate cool-white fluorescent light. There is a finding
as follows: A form of indirect glare which is also known as a veiling reflection was created due
to the light bounced off the surface of that task into employees’ eyes form overhead fixtures. A
change was applied and new selected full spectrum lighting could behave more efficient than the
previous system as it uses less electricity and last longer. Those costs came out to be minor
compared to the productivity improvements after modernization. The new lighting reduced
veiling reflections, which resulted in 13% increase in the employees’ work productivity. The
benefit of improved productivity was estimated as $235,290 per year. Furthermore, a number of
errors in their work were reduced as well. In addition, absenteeism rates seemed to be reduced
after the new lighting systems were installed, and lower eyestrain and fewer headache rates were
reported also. The baseline benefits with better lighting were projected to save $ 255, 929 per
year in PP&L(Deneen 2004). Therefore, using full spectrum lighting could be one way of
improving visual comfort and employee productivity.
Dr. Same Berman, conducted an application of the theory that by enhancing scotopic side of
light (blue light) which was believed to be energy efficient and increase visual accuracy.
(Berman 2000) He reported a lighting experiment results in his article: a demonstration was set
up for the purpose of testing how the new finding on rod sensitivity affected vision and
brightness. “Conventional fluorescent lamps were used to compare the vision effects of a high
color temperature lamp (scotopically enhanced), and therefore higher bluish output, with a low
color temperature lamp (lower bluish output, scotopically deficient)”. Intel facilities staff
observed that under the scotopically rich lighting, people could see better and that this lighting
would be perceived as brighter even though a light meter would display the opposite, which is
exactly what PGE had tried to show. Intel realized that with scotopically rich lamps, they would
25
achieve larger energy saving by reducing the number of lamps and meanwhile maintain or
improve prior vision and brightness conditions. (Berman 2000)
Renowned pediatrician Doris Rapp, MD, stated in her, “Is This Your Child’s World?”(Rapp
1997), that natural light should be the best lighting for not only schools but also anywhere else.
However, in lots of places, students spend more than 6 hours a day under the fluorescent lights
which are in a cool white color. The productivity of students could be decreased because of the
emitted radiation and X-rays from fluorescent lights, which also caused health problems such as
eyestrain, fatigue and depression sometimes among students. A research proved that by replacing
fluorescent lights with full-spectrum lighting would decrease those health problems significantly
which was about 33%. So, full spectrum lighting were encouraged and focused more in both
daily activities and research aspects (Rapp 1997). Based on Dr. Rapp statements, full spectrum
must be chosen if fluorescent lighting must be used. With the recent invented lighting products
with the blue spectrum, it is possible for companies to change lighting environment in the office.

2.3 – Human Pupil Sizes and the Potential Use
Only over the past a few years have researchers discovered the role of pupil size in vision and
importance of designing appropriate interior lighting to maintain visual acuity. Many researches
have investigated for spectrum controls. Scotopically enhanced lamps, which favor the blue-light
wavelength, reduce pupil size. According to Professor Dr. Sam Berman, “At typical interior light
levels, smaller pupils will contribute to better vision. The present lighting practice often calls for
reducing pupil size by raising light levels, which is not efficient and fails to utilize the response
of the rods to control pupil size.” (Berman 2000)
26
In a joint program which was cooperated by Lawrence Berkeley Laboratory and University
of California, San Francisco, human responses to electric lighting were studied. Significant
differences in pupil size occurred when subjects were exposed to indirect high-pressure sodium
(HPS) lighting as compared with indirect incandescent lighting when the light intensities were
photopically matched. The two lighting systems applied roughly the same spatial luminance
distribution. By observing difference in pupil size, the spectral power distribution of the two
lighting systems were claimed to have effect on visual performance and other aspects of visual
systems function.
Pupil size was recognized to have considerable influence on the visual system ability to
achieve higher resolution of details which is also known as visual acuity and on depth of field as
well. (Leibowitz 1952), and spatial contrast sensitivity function (B. Y. F. W. Campbell and
Green 1965). It was observed that depth of field decreased inversely when pupil diameter
increased (F. W. Campbell 1957). It was also noted that larger pupil allowed more retinal
luminance in a steady ambient luminance environment (Luckiesh and Moss 1934; Ferguson and
Stevens 1956). Thus, improvement in visual acuity could be achieved by control of pupil size
independent of light condition depending on the specific factors of the visual task (Eastman and
McNelis 1963). Larger pupil could be realized under a scotopically deficient lamp and was
believed to have better performance than under another lamp which had more scotopic lumens.
On the other hand, studies of contrast sensitivity (B. Y. F. W. Campbell and Green 1965; B.
Y. F. W. Campbell and Gubisch 1966) showed a steady reduction in this quantity with increasing
pupil size. Further studies were encouraged to show that there are preferred pupil sizes in the
everyday world of visual tasks, the results here should lead to a new dimension for improving the
quality of our lighting environment.
27
But given the controversies within vision science, and the importance of pupillary response
to vision and lighting design, further testing on other lighting will be necessary to see how much
effect other factors of lighting and human could have on the pupil. When such additional
information is available, the general principles governing this part of visual efficiency will have
a more ensured base.
There were other researches who investigated environmental and physiological factors
affecting pupil size. It was indicated that pupil size decreased linearly as a function of age at all
illuminance levels. Significant effect of age on pupil size was still shown at the highest
illuminance level. The change rate of pupil diameter caused by age was about 0.043 mm per
year at low illuminance level and about 0.015 per year at high illuminance level. “In addition, the
variability between pupil sizes of subjects of the same age decreased by a factor of
approximately two as luminance was increased over the range investigated. Pupil size was found
to be independent of gender, refractive error, or iris color (P >0.1).”(Winn et al. 1994)
Although those previous studies showed independent of gender, a validation still needs to be
done. Other human subject factors, such as ethnic origin, myopic condition, and age should also
be tested (again) to figure out their potential effects on pupil size.
2.4 – Conclusions for Background Research
The existing researches related to this topic are still very limited. However, the those
researches have already explained the importance of proper lighting conditions in the office
environment which significantly affect employees’ visual comfort, health and productivity and
indicate the potential use of pupil size in the future lighting research. Although there were studies
on environmental and physiological factors affecting human pupil size and there were researches
28
using pupil size for a lighting study. Those studies were mostly focused on improving visual
comfort by controlling spectrum. No one previously studied correlations between illuminance,
luminance and human pupil size and applied it to lighting design and control, which is very
critical in today’s building environment, where people spend more than 90% of their time, for
their visual health and work productivity.

29
Chapter 3: Methodologies
Methodology of the research will be introduced into details in this chapter. First, the scope of
work will be given explaining the main procedure of the research and selected variables; Second,
a description of experimental chamber will be presented to give a whole idea of the environment
where the experiments were conduced; Third, research tools and software used in the research
will also be presented with detailed explanation and index; other preparations including taking
IRB course and preliminary study will be discussed in the end.
3.1 - Scope of Work

Figure 3.1 Conceptual diagram of the research methodologies
The experiment was conducted in a well-designed and equipped chamber, which is based on the
requirements of experiments, on the basement level of Watt Hall on University of Southern
California campus.
30
A desk was placed against the south wall of the room. 16 LED dimmable lamps were
installed above the workstation as the light source that were also controlled by a manual dimmer
for creating different lighting intensity on the desk. Two chairs were placed beside the desk, one
for the participant, the other for student investigator. The chamber was modified or rearranged
based on the experiment plan. The dimmable LED lamps were changed based on demand of
different color temperature. For paper-based task in the experiment, the central part of the desk
was left with enough space for participants’ working, while, for computer-based task, a monitor
plus a keyboard and a mouse were placed on the central section instead.
The chamber has high acoustic insulation and absorption for unfavorable noise, the thermal
condition maintained constant and indoor air quality is ensured by a ventilated fan on the ceiling,
fresh air is provided. A ceiling lighting fixture is installed for general daily work in the
chamber and serves as aided light source.
This is a human subjects experiment, and only one person can participate in the experiment
at one time. There is no specific way in selecting participants. But an effort has been made to
achieve a balanced ratio in human characteristics such as gender, age, myopic condition as well
as eye color. Participant was informed of the whole procedure of the experiment and their main
task as well as some restrictions during the experiment. Each experiment took about 1.5 hours to
conduct. During the entire process, the participant was asked to carry out a typical type of office
work separately using paper-based and computer-based tasks. Every five minutes, the participant
was asked to respond to a series of questions regarding their visual sensations and comfort
conditions. At the same time, pupil size and lighting parametric data were automatically
collected and saved in a computer-based data acquisition server. Detailed information will be
explained in Section 3.3 and 3.4. Based on the collected data of lighting environment conditions,
31
pupil size reactions, and feedback on visual sensation and comfort conditions, a visual comfort
model was developed for future lighting designs and individual control purposes using a
statistical tool.
3.1.1 – Lighting Parameters
Illuminance, luminance and color temperature were selected as indicators for the lighting
conditions and environment in the experimental chamber. As one of basic lighting components
that represent the lighting environment condition, illuminance has been examined in several
studies (Veitch and Newsham 1998; Manav 2007). It is also included in most design guidelines
and as an indicator of energy consumption. Instead of selecting a single visual location for the
data record, overall luminance provides a better understanding of average performance for a
certain area in a typical office environment. Many previous studies have also used luminance as
one of the critical lighting parameters. Color temperature, which has been studied as another
significant lighting parameter having a substantial relationship to the productivity, is also
considered in this project. (Oi and Takahashi)
• Illuminance: “One lumen of luminous flux, uniformly incident on 1 m
2
(ft
2
) of area
produces an illuminance of 1 lux (footcandle [fc]). Illuminance is normally represented
by the letter E. It is the density of luminous power, expressed in terms of lumens per unit
area. If consider a lightbulb as analogous to a sprinkler head, then the rate of water flow
would be the lumens, and the amount of water per unit time per m
2
(ft
2
) of floor area
would be the lux (footcandles). Thus, the SI unit, lux, is smaller than the corresponding
I-P unit, footcandles, by the ratio of square meters to square feet. That is, 10.764 lux =
1fc.” (Grondzik, Kwok, Stein, & Reynolds, 2010, p. 472)
• Luminance: “Luminance is normally defined in terms of intensity; it is the luminous
intensity per unit of apparent (projected) area of a primary (emitting) or secondary
(reflecting) light source. Thus, its units are candela per area. While SI unit of luminance
32
is candela per square meter (cd/m
2
), sometimes referred to as the nit.” (Grondzik, Kwok,
Stein, & Reynolds, 2010, p. 473)
• Color temperature: “The color of the light radiated is related to its temperature. By
developing a blackbody color temperature scale, we can compare the color of a light
source to this scale and assign to it a color temperature – that is, the temperature to which
a blackbody must be heated to radiate a light similar in color to the color of the source in
question. Temperature is measured in Kelvin, which is a scale that has its zero point at
-460
o
F”. (Grondzik, Kwok, Stein, & Reynolds, 2010, p. 514)
Illuminance was measure in lux, luminance was measured in cd/m
2
and color temperature in
Kelvin. The settings were decided based on the consideration of the possible ranges of individual
lighting parameters in office environments in the U.S. (Choi, Loftness, and Aziz 2012). As
typical illuminance level in the office is 400 lux for paper based task. The range from 50 to 1400
lux should cover the preferences of most occupants. For light color temperatures, two conditions
with warm light (2700 K) and daylight (5000 K) were selected. These are the most frequent work
types in a typical workplace. Typical range and interval for the experiment is summarized in
Table 3.1.
Table 3.1 Settings for lighting parameters
Illuminance: 50 – 1400 lux, 150 lux interval
Luminance: Lowest (cd/m
2
): 2.05 (min), 21.23 (max), 8.67 (ave), 14.3 (UGR)
Highest (cd/m
2
): 2.22 (min), 581.61 (max), 278.92 (ave), 10.4 (UGR)
Color temperature: Warm: 2700 K – 3000K; Daylight: 5000K – 6500K

33
3.1.2 – Task Types
Two task types were selected for the human subject experiments: paper-based task and
computer-based tasks. The computer based task consisted of reading and typing. The same
material was supplied to all of the subjects in both digital and printed forms. While, in
paper-based task, the reading material was printed out and provided to participants. All other
settings and experiment procedure were kept same with the computer-based task.
3.1.3 – Questionnaire and Performance Test
A survey questionnaire was used for collecting participants’ identification information and
feedback on their perceived visual sensations and comfort conditions. Collected identification
information included age, gender, ethnicity, eye color and myopic condition. Age, gender and
ethnicity are most common parameters collected and taken into consideration when conducting a
human subject involved research project, especially in the domain of visual comfort studies
(Sivaji et al. 2013). Eye color and myopia are two parameters more directly connected to lighting
studies. All of these parameters are believed to have an effect on physiological responses of
human pupils (Sivaji et al. 2013). Those parameters will be used for grouping the participants for
data analyses to investigate the correlations between physiological and environmental conditions
in this study.
For visual sensation and comfort condition surveys, a 7-point scale was applied to the
answers of the questionnaire. Background research indicated a 7-point scale is better than 5 point
scale to give a higher resolution for answers without excessive complication (Sauro 2014).
Previous studies indicate that applying more than an 11 point scale has a diminishing benefit.
Since there are only 10 change steps of lighting levels generated during the experiment, 11
34
would be too many for the participants. An example of an error or bias that could be introduced
by the scale is that participants might simply pick one option adjacent to the previous one for the
newer level without much considerate thinking or judging. A 7-point scale has been widely used
for thermal comfort research asking for satisfaction levels. This scale consists of both enough
essential options and distinctions between options. It provides neutral level at the central point
and two other directional sections for positive and negative points. And in each section, there is
highly extreme and slightly effective levels and one moderate between them. This type scale
could facilitate participants’ report on their perceived visual sensations and comfort conditions.
An example of the questionnaire shows how it was organized (Figure 3.2).

35

Figure 3.2 Sample Section of Designed Questionnaire
To assess the participants’ visual acuity as their work productivity in each lighting condition,
simplified performance tests were conducted during the experiment with simulating individual’s
typical light office work. Participants were also invited to complete four performance tests at
different lighting levels, including brightest (1400 lux), darkest (50 lux), moderate (800 lux) and
optimal level. The optimal level was selected at the end of each experiment based on their
36
preferred lighting condition. The neutral sensation with the highest satisfaction level was
believed as the optimal level, but it varied depending on the participants.
A selected performance test was a simple reading and typing work. The same reading
material was given in a text file to each participant. Meanwhile, another blank text was open for
typing purpose. These two text files were displayed on the monitor at the same time with same
settings in a font size and background color. The evaluation was conducted in two parts: typing
speed and error rate. Total typed characters were counted as typing speed, and typos were
accumulated to estimate an error rate based on the total character number.
3.1.4 – Pupil Size Parametric Data
As discussed in Chapter 2, the pupil tracking device ASL Mobile Eye-XG was used for
collecting pupil size and its parametric data of human subjects. Each participant was asked to
wear this sensory device for the entire experimental time period (Figure 3.3).

Figure 3.3 Participant Wearing ASL Mobile Eye-XG Device
Due to the presetting defined in the device, the pupil size was measured and collected at 30
Hz, and a very large amount of data was collected from each test. The data process and analysis
procedure were conducted using the Microsoft Excel and Minitab Statistical Package software,
37
and the details would be discussed in Chapter 4. The pupil size was shown in pixels for
illustrating and comparing purposes. Since human pupil sizes vary depending on people, the
study normalized the collected data per individual participant to estimate a changing rate of the
size per lighting condition and compared to those physiological changes between different test
participants or their physiological groups. The normalization process will be discussed in
Chapter 4.
3.2 - Experimental Chamber Setup
An experimental chamber was used for the human subject experiment. It simulated a private
closed office workplace which would be dedicated to only one user without being affected by
other neighbor workplace environmental conditions, and that allowed controlled lighting
conditions. Other indoor environmental quality components, such as thermal, air and acoustic
conditions were monitored and controlled to be maintained at constant levels as only human
physiological responses and lighting environments were being studied.
3.2.1 – Chamber Design
With a financial support from the USC School of Architecture, a room was provided for the
experiment. The room is located on the basement floor of Watt Hall, and the area is about 15 m
2
.
As the windowless room is located on the basement level of a high-mass building, it is not
significantly affected by external thermal and acoustic conditions. The chamber consists of two
interior spaces; one is a test room for a participant, and the other is a monitor room for an
investigator. This physical division helps the participant concentrate on the test without being
distracted.
38
The chamber was composed of two parts: A Test Room and a Monitor Room (Figure 3.4).
For the purpose of reducing other effects on the participant, an individual room was designed for
a participant to take a test while the investigator in stationed in the Monitor Room for
supervising the experiment and offering any help if needed. The division between the rooms has
a large window that occupies more than half of the area to be used so as to allow the
investigator’s view through for a monitoring purpose. Light can leak through the window but
there is no effect on the performance of the participant, since light source is located in the test
room. The interior wall of the chamber was painted plain white, which does not cause any glossy
refection and is a typical interior color in an office environment.

Figure 3.4 Diagrammatic plan of Modifying Chamber
3.2.2 – Lighting Fixtures
There are two types of light sources in the chamber. Fluorescent lamps were originally installed
on the ceiling for the purpose of strengthening the light when needed. The second source is the
primary light source for the experiment. The light source adopted in the experiment is a Philips
dimmable LED lamp. Each lamp provides up to 530 lumens for luminous flux and 2700 K color
39
temperature. For the earlier pilot study – a paper based task, 12 lamps were used. Later, 3 more
lamps were added to achieve a higher lighting illuminance level on the work surface. For the
second round of the experiment, the installed LED light bulbs were replaced with a similar type
of lamps, which emit 5000 K color temperature. Figure 3.5 shows the lamp adopted in the
chamber test.

Figure 3.5 Dimmable LED Lamp
3.3 – Research Tools and Sensor Devices
A detailed description of all research tools and sensor devices used in the research will be given
in this section.
3.3.1 – Sensory Devices: Illuminance Meter, Luminance Meter and HDR camera
Three illuminance meters were used to measure and collect lighting intensity data (Figure 3.6).
They were evenly distributed on the desk so that measure data could reflect more precisely.
Illuminance meters are powered by batteries and could output 0-5V analogue signals to a data
acquisition device (DAQ), which has a functionally featured with a signal conversion to
illuminance value and data storage.
40
The HDR camera (Model: Nikon COOLPIX 8400) (Figure 3.7) with a fisheye lens was also
used for estimating luminance data is. The settings for the camera are summarized in Table 3.2.

Table 3.2 Settings for COOLPIX 8400 in “M” mode
White balance Sunny
Best shot selector Off
Image adjustment Normal
Saturation control Normal
Image quality Normal
Image size 8M (3264 x 2448)
Sensitivity 100 ISO
Image sharpening Off
Lens Fisheye
Exposure option (AE lock) Off
Auto bracketing Off
Noise reduction Off

With the same settings, four photographs for each illuminance level were taken to calculate
luminance levels. Each image was taken with different individual aperture and exposure time
settings, which are summarized in Table 3.3.
41
Table 3.3 Aperture and Exposure Time Settings
COOLPIX 8400
Aperture Exposure Time
3.8 1/2
3.8 1/15
3.8 1/125
3.8 1/1000

All the images taken in each lighting condition, mainly driven by illuminance, were
processed using the Photolux 2.1 software to calculate luminance for each illuminance level.
Highest, lowest, average and UGI were also estimated as representatives for the luminance
performance based on the use of the processed image images.
For a calibration purpose, the Cooke cal-SPOT 401 luminance meter was used (Figure 3.8).
The measured luminance with a high resolution was compared with the estimated by the
Photolux software. The study confirmed that there were minimal discrepancies between the
measured and calculated values.

42

Figure 3.6 OMEGA HHLM-1 Figure 3.7 Coolpix 8400 Figure 3.8 Cooke cal-SPOT 401
3.3.2 –DAQ device
Another essential device adopted in the chamber for data acquisition was the NI USB 6008
(Figure 3.9). It was connected with all the illuminance meters in the chamber and also
transmitting collected data into a laptop through cable. All the data were displayed in real time
on the laptop by a programmed data collection interface developed using in the LabVIEW
software. More detailed information could be figured out in section 3.4. Data were also
automatically saved into a selected file with being labeled with detailed information including
date, time, etc.

Figure 3.9 NI USB-6008
43
3.3.3 –ASL Mobile Eye XG
An essential device is ASL Mobile Eye XG that was used for detecting, measuring and collecting
pupil size data from participants (Figure 3.10). The device consists of three major elements: a
laptop computer with related software installed, monitor and special-designed glasses with
cameras. Three components were connected and working together for pupil size parametric
data acquisition. Participants were asked to wear the sensory glasses during the whole
experiment process. Some adjustments were applied to each participant to achieve the best
display in the laptop. There are two cameras on the glasses, one is for capturing images of the
pupil, while the other is for monitoring the view seen by the human subject. The pupil size data
was automatically saved to the laptop and it can be exported as a csv file for future analysis in
MS Excel and Minitab 16.

Figure 3.10 ASL Mobile Eye XG
(Head Mounted Eye Tracking See the World Through Different Eyes)

44
The final layout of the chamber and locations of all devices are shown in Figure 3.11.

Figure 3.11 Fisheye View of the Chamber

3.4 - Adopted Software
Many software tools were adopted or programmed for this project. A challenge at the beginning
of the project was identifying a program to efficiently collect illuminance data. Later, for the
purpose of luminance analysis and data processing, other software were adopted.
3.4.1 – LabVIEW
LabVIEW is a graphical programming platform that allows users to develop different programs
with desired functions related to data collection, processing and storage. It offers great
compatible connections to many data acquisition hardware which makes it easier to use than
other similar software.
45
LabVIEW software is ideal for any measurement or control system, and the heart of the
National Instrument design platform. Integrating all the tools that engineers and scientists need to
build a wide range of applications in dramatically less time, LabVIEW is a development
environment for problem solving, accelerated productivity, and continual innovation.”
(LabVIEW System Design Software) The best attraction of this software is its significant
advantage in data acquisition researches. It has very good connections to DAQ sensor devices. A
core data collection, processing and storage tool was programmed in LabVIEW platform.
3.4.2 – Programming Logic
Signals transmitted from DAQ sensor are ranging from 0 – 5V. Based on the instructions of
illuminance meter, formula was used and programed to calculate illuminance. The calculated
illuminance was displayed on the laptop. The user can also set intervals and choose desired
channels according to the project purpose. The file name was created using the current date and
time. Figure 3.12 and Figure 3.13 present appearance and graphic programming icons in the
LabVIEW. The interface was modified and developed based on the previous product provided by
Professor Joon-ho Choi.

46

Figure 3.12 Front Panel of Designed Program in LavVIEW

Figure 3.13 Block Diagram of Designed Program in LabVIEW
3.4.3 – Photolux 2.1
Since the luminance meter could only measure one spot at a time, a better way of estimating
luminance level for a specific area would be needed. An HDR camera was selected for this
purpose. In addition, Photolux was adopted for calculating luminance of a specific view of
47
participant by combining all four images taken at different aperture and exposure settings into
single file and conducting analysis (Figure 3.14 and Figure 3.15).

Figure 3.14 Four images taken at different exposure settings for same illuminance setting
displayed in Photolux before combining.

48
Figure 3.15 Processed and analyzed image after combining four images taken at different
exposure settings for same illuminance setting in Photolux.
Four pictures of lighting environment in the chamber were taken under the settings shown in
Table 3.3 of Section 3.3.1. With different aperture and exposure time, the image could have
different reflection of the lighting performance in the chamber. After combing these four images,
a overall explanation of the lighting performance of the chamber could be calculated and shown
in Figure 3.15. This image could comprehensively present the illuminance and luminance of the
chamber.
3.4.4 – Minitab
After the data acquisition process, statistical analysis was applied to datasets to determine
correlations between parameters. Minitab was chosen to finish the objective. Minitab is
a statistics package. It was developed at the Pennsylvania State University by researchers
Barbara F. Ryan, Thomas A. Ryan, Jr., and Brian L. Joiner in 1972. Minitab began as a simple
version of OMNITAB, a statistical analysis software made by NIST; the documentation for
OMNITAB was published 1986, but there has been no apparent development since then.
(OMNITAB 80) Minitab is an easy to learn statistical analysis software. It provides a
considerable functions for statistical analysis including T-test, correlations, regression, ANOVA,
etc. It also has strong graphic generators that help explain and present data better. Figure 3.16
presents a sample interface in Minitab.
49

Figure 3.16 Sample interface of Minitab
3.5 - IRB Preparation
An approval from the USC University Park Institutional Review Board (UPIRB) was required
for human subject research. Meeting requirements outlined in 45 CFR 46.110 category (4), (6)
and (7), the IRB designee determined this research that involves no more than minimal risk.
Requirements were all satisfied. Minors are not eligible for inclusion. Approval of this study was
granted on 3/17/2013.
As the project manager and student investigator, required courses were taken and related
tests were passed so as to be eligible for the experiments. The course completion report is
summarized in Figure 3.17.
50

Figure 3.17 Human Research Curriculum Report
3.6 – Preliminary Study and Results
Before starting actual experiments in the chamber, practice with the devices and software was
done. Pilot studies were conducted in the spring semester of 2013 for the purpose of learning the
system and to practice the methodology of conducting the experiments. A group of 13 students
participated in the pilot study. The demographic information is summarized as follows: eight
females (age: 26.3±2.12) and five males (age: 25.6±1.52), and eight Asians and five Caucasians.
Test methodologies consisted of different settings for illuminance in a paper-based task setting.
In the pilot study, ten different levels of illuminance, following order of lowest to highest,
were tested in the workstation setting of the chamber. The pilot study focuses on investigating
51
the relationship between lighting intensity at the workstation surface and the user’s pupil sizes.
At that time, the room was only equipped with 12 units of 9W-LED lights on the ceiling surface
and data acquisition device, which include lighting sensors, lighting controller, and a computer
for purpose of displaying and collecting data. The generated lighting intensity (lux) on the
workstation surface ranged from 150 to 1050 lux with a 100 lux interval. The test had a
two-minute stand-by period in each lighting level to allow pupil adjust to a new lighting level,
and a one-minute data collection time frame for lighting intensity and pupil size measurement.
The overall range of deviations was approximately ±10 to 25 lux in the test. At the end of each
lighting step, the subject was asked to report the visual sensation using a seven-point scale
questionnaire: (-3) very dark; (-2) dark; (-1) slightly dark; (0) neutral; (+1) slightly bright; (+2)
bright; (+3) very bright. The results for preliminary test will be explained in Chapter 4.
This preliminary test indicated several problems. The first one was that the generated
lighting intensity could not satisfy the desired range. In that case, two more same type dimmable
LED lamps were installed in the chamber that finally could generate more than 1400 lux in the
chamber. The second problem was about a time interval between each illuminance level. Three
minutes total was questioned by some professionals and suggested not enough time for the
human pupil finishing adjusting the new lighting level. For the official experiments, the time
interval was raised from 2-minute standby, 1-minute measurement to 3-minute standby and
2-minute measurement. Empirically based on pilot study, 5 minutes should be enough for
adjustment in this research whose interval is 150 lux between each level. The third problem was
complaints from participants about glare. Because there is a tilted glass fixed on the glasses to
reflect pupil to the camera for tracking purpose, the tilted glass can reflect light from lamps
located above the participant in some case which caused serious glare problems with a result of
52
lower comfort conditions. In order to solve this problem, modifications by installing light-shields
were applied to the chamber at beginning of fall semester 2013.
In order to solve the glare problem, several methods were implemented. First, changing
the location the desk was tested. The seat of the participant was relocated to achieve different
angles avoiding strong reflection of the light source so as to resolve glare problem. The desk was
moved to the opposite side of the chamber, so that the participant would face away from the light
source and thus block the glare. Unfortunately, this resulted in enlarging the distance between
workspace and light source, and the light intensity projected on the desk was reduced below the
acceptable level of 1400 lux.
Later, realizing that the camera installed on the glasses blocked some of the reflection
inspired the idea of a creating a shade for the glasses. A small paper board was attached to the
top of camera; this expanded blocking area and solved the problem. An ordinary baseball-style
cap could also be adopted for the same purpose. Using either the paper board or the cap could
resolve the glare problem and would not affect actual light source from working area.
3.7 – Experiment Rounds
After finishing all preparations, three rounds of experiments were conducted within one year. To
maintain a statistical significance in the study, at least 20 human subjects were sampled per each
round of the designed experiments. The different rounds of the experiments were defined
based on the task-type modes and lighting color temperatures. The settings for three rounds are
summarized in Table 4.
53
Table 3.4 Different Experiment Rounds
Round No. One Two Three
Working Type
Computer based Computer based Paper based
Color Temperature
Warm Daylight Daylight

Since computer based task and daylight are mostly used in office, these two settings account for
2/3 of the whole experiments.
For the demographic characteristics of the participants, the ideal combination would be
balanced populations in ethnicity (e.g. 50% Caucasians and 50% Asians), age groups (younger
than 30 years old and 30 or older), and gender. However, due to the limited environment
condition, the balanced population ratio was not able to be achieved in this research. The
demographic information of the test participants of each round are summarized in Chapter 4.
No food or drink was allowed during the experiment to maintain a consistent physical
condition across the participants. Although caffeine didn’t appear to cause irregular pupil
behavior, each subject was asked not to consume caffeinated food or drink at least one hour
before the experiment so as to avoid any potential effect (Wright et al. 1997),. The participant
was expected to arrive 30 minutes earlier before experiment starting to remain in a stable
metabolic rate. A detailed explanation about the process was given to the participant and
questions were allowed to ask if there were any. Most of conversations between participant and
investigator were minimized to avoid any erroneous data collection, such as facial expression
54
changes affecting pupil size detection. Necessary time alerts and reminders were offered to the
participants during the experiment. The participant was seated at the table with the classes and
asked to perform the tasks (Figure 20).

Figure 3.18 Images taken for some participants in the experiment
3.8 – DATA Analysis
When each round accomplished, summarizing collected data and analyzing data was applied.
The parametric data about pupil size, illuminance, luminance, responses to questions and subject
personal information were processed and combined into a single dataset for each participant that
resulted in a tab in the Excel spreadsheet. Figure 21 shows the sample dataset in Excel
spreadsheet. Minitab was used for statistical analysis. The results and analysis would be
discussed in Chapter 4.
55

Figure 3.19 Data Analysis Process

This chapter presents detailed information about the scope of the research, experimental chamber,
research tools, adopted software and other preparation and procedure of the project. By being
acquainted with the methodology of this research, it would be easier to understand the results and
discussion coming in the next chapters.

56
Chapter 4: Study Results
After a pilot study and following the completion of each round of experiments, all data were
collected for analysis. In this chapter, the pilot study results are summarized and briefly
discussed. Then, data is presented about each individual’s preferred range of lighting
environment and their pupil size based on the visual sensation levels. The overall pupil size
range of all the subjects in each round of chamber experiments is also summarized. The
displayed data is all raw data without any processing such as standardization.
4.1 – Pilot Study Results
Figure 4.1 summarizes the ranges of illuminance per visual sensation for all individual study
participants. Volunteer subjects showed different levels or ranges of illuminance with the same
sensation level. For example, the subject #01 illustrated around 400 lux as a “neutral” level while
illuminance between 200 and 700 lux were reported as a “slightly dark” or “slightly bright”
condition by the test participants. However, the subject #03 reported 170 lux as a “neutral”
condition (i.e. satisfied), but the range from 250 lux to 1000 was “(slightly) bright” for the
subject. The data in Figure 4.2 confirms individually different preferences on illuminance.
Illuminance ranges perceived by the test subjects were aggregated at each visual sensation level
in the figure. The illuminance range of the subjects’ neutral sensation widely ranged from 140
lux to 700 lux, and “bright” sensations were reported with conditions between 250 lux and 1000
lux.
57

Figure 4.1 Illuminance range per visual sensation of each individual (Pilot study).
Subject No.
Sensation
#13
#12
#11
#10
#09
#08
#07
#06
#05
#04
#03
#02
#01
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
1200
1000
800
600
400
200
0
Illuminance (lux)
58

Figure 4.2 Overall illuminance distribution per visual sensation of all individuals (Pilot study).
Measured pupil sizes also showed various features per individual test subject. Some subjects
showed relatively small ranges of pupil size changes between the visual sensations, but others
generated large variations in their pupil sizes across the visual sensations. Currently, the only
reason could be came up is individual physiological characteristics. To reduce the deviations by
personal physical conditions, the collected pupil size data were standardized per individual
subject by using the following equation:

where i is a visual sensation.
Individually standardized values by the equation showed more stability with reduced
deviations across the subjects (Figure 4.3) than the raw data of pupil sizes. As shown in Figure
980 840 700 560 420 280 140
-2
-1
0
1
2
3
Illuminance (lux)
Visual Sensation
100
) _ ( _
) _ ( _ ) ( _
(%) _ _ tan ×
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛ −
=
sensation neutral size Pupil
sensation neutral size Pupil i size Pupil
size Pupil dardized S
59
4.3, the confidence intervals of each subject show more consistency between the different
sensations across the test subjects. In addition, the standardized data showed clear ranges of
confidence intervals between the different sensations. These results illustrate the potential of
pupil sizes to characterize visual sensation.

Figure 4.3 Ranges of standardized pupil size per visual sensation of each individual (Pilot study).
Subject No.
Sensation
#13
#12
#11
#10
#09
#08
#07
#06
#05
#04
#03
#02
#01
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
125.0%
100.0%
75.0%
50.0%
25.0%
0.0%
-25.0%
-50.0%
Standardized Pupil Size
60

Figure 4.4 Overall standardized pupil size distribution per visual sensation of all individuals
(Pilot study).

The standardized pupil sizes per individual subject were combined to find a common feature
across the subjects. Overall distributions of standardized data in Figure 4.4 are clearly shaped
like a typical normal distribution in each visual sensation, and the pupil size ranges seem
distinctly differentiated from each other. The ANOVA (analysis of variance) test in Table 4.1
(left) shows that those distributions are clearly set apart from each other with a statistically
significant p-value of 0.000.
The confidence interval (CI) of the variance of standardized pupil size distribution of each
visual sensation level is evidently separated from each other, so it is enough to show a visual
sensation by reading the ranges of the CIs (Table 4.1- left). A stepwise regression formula also
shows the significant contribution of standardized pupil size to predict the visual sensation
(Table 4.1- right) with a significant p-value, and a R-sq value increase in the stepwise regression.
90.0% 60.0% 30.0% 0.0% -30.0% -60.0% -90.0%
-2
-1
0
1
2
3
Standardized pupil size (%)
Visual Sensation
61
The stepwise regression generated a high R-sq of 70.25, which indicates a higher predictive
potential increase of standardized size than the actual pupil size adopted in the step 3. Therefore,
standardized pupil size can be used as a critical single predictor to estimate an actual visual
sensation.

Table 4.1 Results of ANOVA with standardized pupil size data (left) and stepwise regression
based on the data of standardized pupil size, actual pupil size and illuminance.
Source DF SS MS F P
Sensation 5 117.3903 23.4781 729.80 0.000
Error 5452 175.3949 0.0322
Total 5457 292.7853

S = 0.1794 R-Sq = 40.09% R-Sq(adj) = 40.04%

Individual 95% CIs For Mean Based on
Pooled StDev
Level N Mean StDev ---+---------+---------+---------+-
-2 30 0.3504 0.1078 (---*--)
-1 747 0.1739 0.1900 (*
0 1033 0.0126 0.1322 (*
1 1555 -0.1362 0.1959 *)
2 1908 -0.2437 0.1860 (*
3 185 -0.1645 0.1554 (-*)
---+---------+---------+---------+--
-0.20 0.00 0.20 0.40
Step 1 2 3
Constant -0.9174 -0.7539 -1.7840

Illuminance (lux) 0.00326 0.00280 0.00294
T-Value 106.30 78.19 81.89
P-Value 0.000 0.000 0.000

Pupil size (Stand.) -0.994 -1.555
T-Value -22.70 -28.70
P-Value 0.000 0.000

Pupil size (Abs.) 0.01435
T-Value 16.83
P-Value 0.000

S 0.643 0.614 0.599
R-Sq 67.44 70.25 71.72
R-Sq(adj) 67.43 70.24 71.70
Mallows Cp 825.0 285.2 4.0

4.2 – First Round Results: Low light color temperature condition, Computer task type
As mentioned in Chapter 3 Methodology, color temperature used for the first round was 2700 K,
which was included in the “warm” temperature range, and the selected work type was a
computer-based task. The remaining variables for the experiment were maintained consistently
for each experiment in the different rounds. The overall ranges of the air temperature, relative
humidity, and CO2 during the experiment were 23.5± 0.7℃, 33±2.5%, and 620 ± 35ppm,
respectively. 20 volunteers participated in the first round experiments. Information about the
human subjects is summarized in Table 4.2.
62
Table 4.2 Demographic information of human subjects (First round)
First Round (Computer + Warm)
Gender Age Eye color Myopic
Male Female <25 >=25 Blue Brown Yes No
12 8 11 9 4 16 9 11
60% 40% 55% 45% 20% 80% 45% 55%

Similar procedures applied in the pilot study were adopted for displaying the collected raw
data. First, the preferred range of illuminance per visual sensation of each participant in the first
round is all displayed (Figure 4.5). Although the all subjects show very different patterns, they
still share a similar general trend: a sensation increases when illuminance increases. That also
demonstrates the basic physiological ability of perceiving the light and visual acuity performance.
Furthermore, several subjects have even very similar patterns, e.g. Subject no. 10, 11 and 19.
That creates the possibility of grouping people for further analyses to determine the similarities
that could commonly found in a certain subject group defined in this research.
63

Figure 4.5 Illuminance range per visual sensation of each individual (First round).
One more step into the analysis of raw data about preferred range for each sensation was
conducted (Figure 4.6). Overall preferred illuminance distribution per each sensation is shown.
Based on the distributions, the general preferred range could be identified. For -3, it lasts from 0
to 300 lux; for -2 and -1, it lasts from 0 to about 1000; for 0, it is from 0 to 1600; for 1 and 2, it is
from 300 to 1600, while for 3, it is 1200 from 1500. There are some data points beyond the
designated highest illuminance level, which is 1400 lux. It was because some extra electric
current caused in the transmission between illuminance meter and DAQ sensor. But these data
points are very few compared to the remaining amount of data and thus are not expected to have
a significant effect on the results and analysis, which will be discussed in Chapter 5.
No.
Sensation
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
1800
1600
1400
1200
1000
800
600
400
200
0
Ave Illuminance (lux)
Boxplot of Ave Illuminance
64

Figure 4.6 Overall illuminance distribution per visual sensation of all individuals (First round).
The pupil size of each participant was collected and totally 1200 pupil size data per person
was recorded for a one-hour experiment. Ranges of original pupil size per visual sensation of
each individual are plotted (Figure 4.7). Due to physiological characteristics of each individual,
they have wide ranges of pupil sizes even in a same visual sensation, No. 6 shows 30 to 80 pixels
in the pupil sizes, while No.1 generates around 40 to 50. Due to individually different pupil sizes
and their change rates, it is difficult to find a generalized pattern among the test subjects.
However, at least all individuals still share similar trends in patterns. Pupil size decreases when
sensation increases, which could also be recognized in daily time.
1800 1500 1200 900 600 300 0
-3
-2
-1
0
1
2
3
Ave Illuminance (lux)
Sensation
Dotplot of Ave Illuminance
Each symbol represents up to 33 observations.
65

Figure 4.7 Ranges of original pupil size per visual sensation of each individual (First round).
The original pupil size of all individuals were combined and plotted (Figure 4.8). Pupil size
doesn’t follow a normal distribution per visual sensation since the physiological difference
among human subjects, which indicates the importance of standardization process of the pupil
size.
No.
Sensation
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
100
90
80
70
60
50
40
30
20
Pupil Size (pixels)
Boxplot of Pupil Size
66

Figure 4.8 Overall original pupil size distribution per visual sensation of all individuals (First
round).

4.3 – Second Round Results: High light color temperature condition, Computer task type
The color temperature used for the second round was 5000 K, which was included in the
“daylight” defined range, and the selected work type was a computer-based task. Compared with
the first round, other variables for the experiment were kept same for each experiment, except
the lighting color temperature in the second round test. Again, 20 volunteers participated in the
first round experiments. Information about the human subjects is summarized in Table 4.3.

100 80 60 40 20 0
-3
-2
-1
0
1
2
3
Pupil Size (pixels)
Sensation
Each symbol represents up to 29 observations.
Dotplot of Pupil Size
67
Table 4.3 Demographic information of human subjects (First round)
Second Round (Computer + Daylight)
Gender Age Eye color Myopic
Male Female <25 >=25 Blue Brown Yes No
12 8 8 12 5 15 7 13
60% 40% 40% 60% 25% 75% 35% 65%

Same as the first round, boxplots of illuminance range per visual sensation of each
individual (Figure 4.9), overall illuminance distribution per visual sensation of all individuals
(Figure 4.10), pupil size range for each individual (Figure 4.11) and overall pupil size
distribution of all individuals (Figure 4.12) were generated based on the experiment data of the
second round. Compared to the findings in the first round experiments, similar findings about
illuminance-sensation patterns and pupil size-sensation pattern were revealed. The raw data are
summarized in Figures 4.9 to 4.12.
68

Figure 4.9 Illuminance range per visual sensation of each individual (Second round).

No.
Sensation
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
1800
1600
1400
1200
1000
800
600
400
200
0
Ave Illuminance (lux)
Boxplot of Ave Illuminance
1800 1500 1200 900 600 300 0
-2
-1
0
1
2
3
Ave Illuminance (lux)
Sensation
Dotplot of Ave Illuminance
Each symbol represents up to 36 observations.
69
Figure 4.10 Overall illuminance distribution per visual sensation of all individuals (Second
round).

Figure 4.11 Ranges of original pupil size per visual sensation of each individual (Second round).

No.
Sensation
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
90
80
70
60
50
40
30
Pupil Size (pixels)
Boxplot of Pupil Size
70

Figure 4.12 Overall original pupil size distribution per visual sensation of all individuals (Second
round).

4.4 – Third Round Results: High light color temperature condition, Paper task type
The color temperature used for the third round was 5000 K, which was included in the “daylight”
defined range. However, the selected work type was a paper-based task. Other variables for the
experiment except the task type were kept same in the third round experiment, compared to the
second round. 20 volunteers participated in this round. Information about the human subjects is
summarized in Table 4.4.

100 80 60 40 20 0
-2
-1
0
1
2
3
Pupil Size (pixels)
Sensation
Each symbol represents up to 30 observations.
Dotplot of Pupil Size
71
Table 4.4 Demographic information of human subjects (Third round)
Third Round (Paper + Daylight)
Gender Age Eye color Myopic
Male Female <25 >=25 Blue Brown Yes No
11 9 7 13 6 14 9 11
55% 45% 35% 65% 30% 70% 45% 55%

In a same way adopted in the previous rounds, boxplots of illuminance range per visual
sensation of each individual (Figure 4.13), overall illuminance distribution per visual sensation
of all individuals (Figure 4.14), pupil size range for each individual (Figure 4.15) and overall
pupil size distribution of all individuals (Figure 4.16) were generated based on the experiment
data of second round. Similar findings about trend of illuminance-sensation pattern and pupil
size-sensation pattern were revealed compared to the first and second found experiments. The
data distributions are summarized in the following figures.
72

Figure 4.13 Illuminance range per visual sensation of each individual (Third round).

Figure 4.14 Overall illuminance distribution per visual sensation of all individuals (Third round).

No.
Sensation
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
1800
1600
1400
1200
1000
800
600
400
200
0
Ave Illuminance (lux)
Boxplot of Ave Illuminance
1500 1250 1000 750 500 250 0
-3
-2
-1
0
1
2
3
Ave Illuminance (lux)
Sensation
Dotplot of Ave Illuminance
Each symbol represents up to 38 observations.
73

Figure 4.15 Ranges of original pupil size per visual sensation of each individual (Third round).

Figure 4.16 Overall original pupil size distribution per visual sensation of all individuals (Third
round).
No.
Sensation
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
80
70
60
50
40
30
Pupil Size (pixels)
Boxplot of Pupil Size
100 80 60 40 20 0
-3
-2
-1
0
1
2
3
Pupil Size (pixels)
Sensation
Each symbol represents up to 27 observations.
Dotplot of Pupil Size
74
Chapter 5: Data Analysis and Discussion
The pupilometer used the pixel as a metric for measurement. It detected the size of a pupil by the
micro-camera facing the subject’s eye while tracking the path of eye movement. The raw data of
individuals’ pupil sizes are not comparable because pupil sizes and shapes vary in different
individuals. For this reason, normalized (i.e., standardized) data for each individual was used for
data analysis using the formula introduced in the pilot study part.
(5-1)
where i is an eye’s response to illuminance.

5.1 – First Round
Figure 5.1 displays the standardized pupil size for each human subject based on their sensations.
The individually normalized data shows more stable fluctuations than the raw data pupil size
data. As illustrated in the standardization formula above, the pupil size measured at the neutral
visual sensation was selected as a baseline. This process relatively flattened the undulation of the
measured data per individual. As illustrated in Figure 5.1, most test subjects showed positive
change rates with darker perceptions and vice versa with brighter perceptions. However, in the
case of No.11, since he reported a “neutral” sensation for the highest illuminance condition, the
measure pupil sizes had increased to 60% of the baseline pupil size. On the other hand, Subject
No. 20 reported a neutral sensation for the lowest illuminance condition, and the measured pupil
sizes had decreased to 25% of the subject’s baseline pupil size. In addition, the normalized pupil
sizes showed different changing rates per individual and his/her visual sensation. Subject No. 5
Standardized_Pupil_size(%)=(
Pupil_size(i)−Pupil_size(neutral_sensation)
Pupil_size(neutral_sensation)
)×100
75
and No. 20 showed very mild change slopes as the visual sensations increased, but data for
Subjects No. 11 and No. 19 presented rapid changes with irregular patterns.

Figure 5.1 Standardized pupil size distribution in each subject’s test (“No.” indicates a subject ID)
(First round).
Figure 5.2 illustrated the pupil size patterns for visual sensations based on the combined data
of all individuals. Overall, the standardized pupil sizes decreased while the generated
illuminance intensity was increasing. The analysis of variance (ANOVA) test showed a
statistically significant p-value that was lower than 0.05 (Table 5.1). This finding is clearly
summarized in Figure 5.3. The chart contains basically the same data as Figure 5.2, but it shows
No.
Sensation
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
100%
75%
50%
25%
0%
-25%
-50%
Individually Normalized Pupil Size (%)
Boxplot of Normalized Pupil Size
76
a 95% confidence interval for pupil sizes per visual sensation. The interval lines are clearly
differentiated from each other, and the length of an interval at neutral sensation is shortest, which
indicates that the pupil size for a neutral sensation is more stable than for other sensations.

Figure 5.2 Overall standardized pupil size distribution per visual sensation to illuminance
intensity (first round).

40% 20% 0% -20% -40%
-3
-2
-1
0
1
2
3
Individually Normalized Pupil Size (%)
Sensation
Each symbol represents up to 29 observations.
Dotplot of Normalized Pupil Size
77

Figure 5.3 Interval plot of standardized pupil size per visual sensation to illuminance intensity
(First round).
However, the mean value of sensation 3 is a bit higher than sensation 2, which could
indicate a lack of enough data for this level. It is also shown in Figure 5.2, which the distribution
of sensation 3 is flat and there is limited number of dots. Similarly, for both sensation 3 and
sensation -3, the interval lines are longer than others because of less reported data from
experiments. That is also an indication that in the current illuminance range (50 lux – 1400 lux),
very few participants considered “very dark” or “very bright.” Another finding is the differences
between sensations when reported among “dark” zones are larger than those among “bright”
zones.
3 2 1 0 -1 -2 -3
20%
15%
10%
5%
0%
-5%
-10%
Sensation
Individually Normalized Pupil Size (%)
95% CI for the Mean
Interval Plot of Normalized Pupil Size
78

Figure 5.4 Boxplot of standardized pupil size per visual sensation to illuminance intensity (First
round).
In addition to Figure 5.2 and Figure 5.3 for the illustration of standardized pupil size
distribution, all data was also presented in boxplot chart (Figure 5.4). The line in the boxplot
indicating median value for each sensation follows similar pattern as mean value has in Figure
5.3.

3 2 1 0 -1 -2 -3
60%
50%
40%
30%
20%
10%
0%
-10%
-20%
-30%
Sensation
Individually Normalized Pupil Size (%)
Boxplot of Normalized Pupil Size
79
Table 5.1 One-way ANOVA test: Standardized Pupil Size versus Sensation
Source DF SS MS F P
Sensation 6 59.7410 9.9568 957.99 0.000
Error 19680 204.5442 0.0104
Total 19686 264.2852

S = 0.1019 R-Sq = 22.60% R-Sq(adj) = 22.58%

Individual 95% CIs For Mean Based on Pooled StDev
Level N Mean StDev -+---------+---------+---------+--------
-3 353 0.1678 0.2029 (-*)
-2 1706 0.1013 0.1405 *)
-1 3909 0.0441 0.0994 *)
0 6216 -0.0001 0.0693 *
1 3120 -0.0198 0.1100 *)
2 4073 -0.0741 0.1091 *)
3 310 -0.0607 0.0737 (*-)
-+---------+---------+---------+--------
-0.070 0.000 0.070 0.140

Pooled StDev = 0.1019

To check the consistency of pupil size changes per visual sensations for individuals, the
study conducted comparison tests between subject groups of different physiological
characteristics, (i.e., eye color, age, gender, and myopic conditions). Since the study adopted 20
human subjects for the chamber experiments, the data for individual sensations were regrouped
from a 7-point scale to a 3-point scale to keep the scope of data at a level for statistical
significance in the data analysis. Therefore, visual sensations of -3 (very dark), -2 (dark) and -1
(slightly dark) were grouped into “dark,” and visual sensations of +1 (slightly bright), +2 (bright)
and +3 (very bright) were grouped into “bright.”
Figure 5.5 shows the interval plot of normalized pupil size with a 95% confidence interval
for the mean value in eye color groups. Figure 5.6 indicates four quartile distributions of data and
comparison sets illustrate a similar pattern between two eye color groups. Each eye color group
80
showed larger normalized pupil sizes at the dark sensation and smaller pupil sizes at the bright
sensation. Since the neutral sensation was reported in different illuminance levels, depending on
the participants, the pupil sizes were also measured in some ranges at the neutral condition.

Figure 5.5 Interval plot of comparisons of overall standardized pupil size per visual sensation
between eye color groups (first round).

Eye Color
Sensation_1
Brown Blue
3_Bright 2_Neutral 1_Dark 3_Bright 2_Neutral 1_Dark
15%
10%
5%
0%
-5%
Individually Normalized Pupil Size (%)
95% CI for the Mean
Interval Plot of Normalized Pupil Size
81

Figure 5.6 Boxplot of comparisons of overall standardized pupil size per visual sensation
between eye color groups (first round).
The analysis of variance (ANOVA) test (Table 5.2) reported significant differences of pupil
size in each group. The comparisons presented that average pupil size change compared with
natural level was about 13.6% at dark and -6.2% at bright in the blue eye color group and
respectively 5.1% and -4.7% in the brown eye color group. The ANOVA test showed a p-value
of 0.000, which is smaller than 0.05 in the level of 95% confidence. These statistical findings
support the concept that the visual sensations can be matched with normalized pupil sizes across
the test subjects, and the findings are also commonly applicable to the eye color groups.

Eye Color
Sensation_1
Brown Blue
3_Bright 2_Neutral 1_Dark 3_Bright 2_Neutral 1_Dark
40%
30%
20%
10%
0%
-10%
-20%
-30%
Individually Normalized Pupil Size (%)
Boxplot of Normalized Pupil Size
82
Table 5.2 One-way ANOVA test: Standardized Pupil Size versus Sensation between Eye Colors
ANOVA for Blue ANOVA for Brown
Source DF SS MS F
P
Sensation_1 2 28.66370 14.33185 1553.16
0.000
Error 4130 38.10969 0.00923
Total 4132 66.77339

S = 0.09606 R-Sq = 42.93% R-Sq(adj) = 42.90%

Level N Mean StDev
1_Dark 1171 0.13628 0.13431
2_Neutral 1040 0.00000 0.05357
3_Bright 1922 -0.06172 0.08544

Source DF SS MS F
P
Sensation_1 2 24.9154 12.4577 1125.10
0.000
Error 15551 172.1886 0.0111
Total 15553 197.1040

S = 0.1052 R-Sq = 12.64% R-Sq(adj) = 12.63%

Level N Mean StDev
1_Dark 4797 0.0510 0.1177
2_Neutral 5176 -0.0001 0.0721
3_Bright 5581 -0.0473 0.1188

Individual 95% CIs For Mean Based on Pooled StDev
Level -+---------+---------+---------+--------
1_Dark (*)
2_Neutral (*)
3_Bright (*
-+---------+---------+---------+--------
-0.060 0.000 0.060
0.120

Pooled StDev = 0.09606

Individual 95% CIs For Mean Based on Pooled StDev
Level -------+---------+---------+---------+--
1_Dark (*)
2_Neutral (*)
3_Bright (*)
-------+---------+---------+---------+--
-0.030 0.000 0.030
0.060

Pooled StDev = 0.1052

The subjects were also grouped by age for the comparison of pupil sizes. Figure 5.7 and
Figure 5.8 illustrate the changing pattern in normalized pupil size at various visual sensations per
age group. The Illuminating Engineering Society of North America (IESNA) has also
categorized human ages into three groups for recommending different levels of illuminance:
younger than 25; 25 to younger than 45, and 45 or older. Therefore, since the range of the test
subjects’ ages were from 19 to 40, the subjects were divided into two groups: junior and senior,
based on the age 25 as a threshold and considering a sample size balance between the groups.

83

Figure 5.7 Interval plot of comparisons of overall standardized pupil size per visual sensation
between age groups (first round).
Age Group
Sensation_1
Senior Junior
3_Bright 2_Neutral 1_Dark 3_Bright 2_Neutral 1_Dark
10%
8%
5%
3%
0%
-3%
-5%
Individually Normalized Pupil Size (%)
95% CI for the Mean
Interval Plot of Normalized Pupil Size
84

Figure 5.8 Boxplot of comparisons of overall standardized pupil size per visual sensation
between age groups (first round)
The average pupil size change compared with natural level was about 7.7% at dark and -5.6%
at bright in the Junior group and respectively 6.0% and -4.5% in the Senior group. The ANOVA
test (Table 5.3) showed a p-value of 0.000, which is smaller than 0.05 in the level of 95%
confidence. These statistical findings support the concept that the visual sensations can be
matched with normalized pupil sizes across the test subjects, and the findings are also commonly
applicable to both the junior and senior groups.

Age Group
Sensation_1
Senior Junior
3_Bright 2_Neutral 1_Dark 3_Bright 2_Neutral 1_Dark
40%
30%
20%
10%
0%
-10%
-20%
-30%
Individually Normalized Pupil Size (%)
Boxplot of Normalized Pupil Size
85
Table 5.3 One-way ANOVA test: Standardized Pupil Size versus Sensation between Age Groups
ANOVA for Senior ANOVA for Junior
Source DF SS MS F
P
Sensation_1 2 18.5696 9.2848 845.42 0.000
Error 8878 97.5027 0.0110
Total 8880 116.0723

S = 0.1048 R-Sq = 16.00% R-Sq(adj) = 15.98%

Level N Mean StDev
1_Dark 3218 0.0598 0.1242
2_Neutral 2176 -0.0002 0.0789
3_Bright 3487 -0.0454 0.0993

Source DF SS MS F
P
Sensation_1 2 28.7980 14.3990 1303.89
0.000
Error 10803 119.2991 0.0110
Total 10805 148.0971

S = 0.1051 R-Sq = 19.45% R-Sq(adj) = 19.43%

Level N Mean StDev
1_Dark 2750 0.0770 0.1270
2_Neutral 4040 -0.0000 0.0636
3_Bright 4016 -0.0558 0.1208

Individual 95% CIs For Mean Based on Pooled StDev
Level ------+---------+---------+---------+---
1_Dark
(*)
2_Neutral (-*)
3_Bright (*)
------+---------+---------+---------+---
-0.030 0.000 0.030 0.060

Pooled StDev = 0.1048

Individual 95% CIs For Mean Based on Pooled StDev
Level -------+---------+---------+---------+--
1_Dark
(*)
2_Neutral (*)
3_Bright (*)
-------+---------+---------+---------+--
-0.035 0.000 0.035
0.070

Pooled StDev = 0.1051

The myopic condition was also selected for grouping the test subjects. The pupilometer,
adopted for pupil size measurement in the study, has two wearable features. The embedded
camera to face the user’s eye could be re-attached to the subject’s sunglasses frame from the
original pupilometer goggle frame.
86

Figure 5.9 Interval plot of comparisons of overall standardized pupil size per visual sensation
between myopic groups (first round).

Myopic
Sensation_1
Y N
3_Bright 2_Neutral 1_Dark 3_Bright 2_Neutral 1_Dark
13%
10%
8%
5%
3%
0%
-3%
-5%
Individually Normalized Pupil Size (%)
95% CI for the Mean
Interval Plot of Normalized Pupil Size
87

Figure 5.10 Boxplot of comparisons of overall standardized pupil size per visual sensation
between myopic groups (First round)
Figure 5.9 and Figure 5.10 show the comparison between the myopic (“Y”) and non-myopic
(“N”) groups. The average pupil size change compared with natural level was about 12% at dark
and –5.2% at bright in the “N” group and respectively 1.8% and -5.0% in the “Y” group. The
ANOVA test (Table 5.4) showed a p-value of 0.000, which is smaller than 0.05 in the level of 95%
confidence. These statistical findings support the concept that the visual sensations can be
matched with normalized pupil sizes across the test subjects, and the findings are also commonly
applicable to both the myopic condition groups.

Myopic
Sensation_1
Y N
3_Bright 2_Neutral 1_Dark 3_Bright 2_Neutral 1_Dark
40%
30%
20%
10%
0%
-10%
-20%
-30%
-40%
Individually Normalized Pupil Size (%)
Boxplot of Normalized Pupil Size
88
Table 5.4 One-way ANOVA test: Standardized Pupil Size versus Sensation between Myopic
Groups
ANOVA for N ANOVA for Y
Source DF SS MS F
P
Sensation_1 2 52.9331 26.4666 2201.96
0.000
Error 11052 132.8398 0.0120
Total 11054 185.7729

S = 0.1096 R-Sq = 28.49% R-Sq(adj) = 28.48%

Level N Mean StDev
1_Dark 3057 0.1155 0.1264
2_Neutral 3110 -0.0000 0.0642
3_Bright 4888 -0.0517 0.1207

Source DF SS MS F
P
Sensation_1 2 6.60785 3.30393 405.61
0.000
Error 8629 70.28856 0.00815
Total 8631 76.89641

S = 0.09025 R-Sq = 8.59% R-Sq(adj) = 8.57%

Level N Mean StDev
1_Dark 2911 0.01755 0.10363
2_Neutral 3106 -0.00017 0.07414
3_Bright 2615 -0.04961 0.09168

Individual 95% CIs For Mean Based on Pooled StDev
Level -+---------+---------+---------+--------
1_Dark (*)
2_Neutral (*)
3_Bright (*
-+---------+---------+---------+--------
-0.050 0.000 0.050 0.100

Pooled StDev = 0.1096

Individual 95% CIs For Mean Based on Pooled StDev
Level -------+---------+---------+---------+--
1_Dark
(-*)
2_Neutral (-*-)
3_Bright (-*-)
-------+---------+---------+---------+--
-0.040 -0.020 0.000
0.020

Pooled StDev = 0.09025

The data was also grouped by genders and was shown in Figure 5.11 and Figure 5.12. The
average pupil size change compared with natural level was about 6.8% at dark and –5.3% at
bright in the female group and respectively 6.8% and -4.8% in the male group. The ANOVA test
(Table 5.5) showed a p-value of 0.000, which is smaller than 0.05 in the level of 95% confidence.
These statistical findings support the concept that the visual sensations can be matched with
normalized pupil sizes across the test subjects, and the findings are also commonly applicable to
both the myopic condition groups.
89

Figure 5.11 Interval plot of comparisons of overall standardized pupil size per visual sensation
between gender groups (first round).

Gender
Sensation_1
Male Female
3_Bright 2_Neutral 1_Dark 3_Bright 2_Neutral 1_Dark
8%
5%
3%
0%
-3%
-5%
Individually Normalized Pupil Size (%)
95% CI for the Mean
Interval Plot of Normalized Pupil Size
90

Figure 5.12 Boxplot of comparisons of overall standardized pupil size per visual sensation
between gender groups (first round)
Table 5.5 One-way ANOVA test: Standardized Pupil Size versus Sensation between Gender
Groups
ANOVA for Female ANOVA for Male
Source DF SS MS F
P
Sensation_1 2 16.8548 8.4274 658.09 0.000
Error 7767 99.4629 0.0128
Total 7769 116.3177

S = 0.1132 R-Sq = 14.49% R-Sq(adj) = 14.47%

Level N Mean StDev
1_Dark 1451 0.0683 0.1284
2_Neutral 1859 -0.0000 0.0756
3_Bright 4460 -0.0530 0.1207
Source DF SS MS F
P
Sensation_1 2 25.5214 12.7607 1289.11
0.000
Error 11914 117.9353 0.0099
Total 11916 143.4568

S = 0.09949 R-Sq = 17.79% R-Sq(adj) = 17.78%

Level N Mean StDev
1_Dark 4517 0.06756 0.12495
2_Neutral 4357 -0.00012 0.06648
Gender
Sensation_1
Male Female
3_Bright 2_Neutral 1_Dark 3_Bright 2_Neutral 1_Dark
40%
30%
20%
10%
0%
-10%
-20%
-30%
-40%
Individually Normalized Pupil Size (%)
Boxplot of Normalized Pupil Size
91
3_Bright 3043 -0.04801 0.09624

Individual 95% CIs For Mean Based on Pooled StDev
Level ------+---------+---------+---------+---
1_Dark
(-*)
2_Neutral (*)
3_Bright (*)
------+---------+---------+---------+---
-0.035 0.000 0.035
0.070

Pooled StDev = 0.1132

Individual 95% CIs For Mean Based on Pooled StDev
Level -----+---------+---------+---------+----
1_Dark
(*)
2_Neutral (*)
3_Bright (*)
-----+---------+---------+---------+----
-0.035 0.000 0.035
0.070

Pooled StDev = 0.09949

5.2 – Second Round
Second round adopted same procedure of data analysis with first round. Standardization of pupil
size and grouping sensations into “Dark”, “Neutral” and “Bright” based on physiological
features were applied.
92

Figure 5.13 Standardized pupil size distribution in each subject’s test (“No.” indicates a subject
ID) (second round).

No.
Sensation
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
3
2
1
0
-1
-2
75%
50%
25%
0%
-25%
-50%
Individually Normalized Pupil Size (%)
Boxplot of Normalized Pupil Size
93

Figure 5.14 Overall standardized pupil size distribution per visual sensation to illuminance
intensity (second round).

40% 20% 0% -20% -40%
-2
-1
0
1
2
3
Individually Normalized Pupil Size (%)
Sensation
Each symbol represents up to 26 observations.
Dotplot of Normalized Pupil Size
94

Figure 5.15 Interval plot of standardized pupil size per visual sensation to illuminance intensity
(second round).
The pupil size patterns for visual sensations based on the combined data of all individuals are
shown in Figure 5.14. Overall, the standardized pupil sizes decreased while the generated
illuminance intensity was increasing. Figure 5.15 contains basically the same data as Figure 5.14,
but it shows a 95% confidence interval for pupil sizes per visual sensation. Except sensation -2
and sensation -1, the interval lines are clearly differentiated from each other. There was no
participant reporting sensation -3 (“very dark”) in the second round.

3 2 1 0 -1 -2
10%
5%
0%
-5%
-10%
-15%
Sensation
Individually Normalized Pupil Size (%)
95% CI for the Mean
Interval Plot of Normalized Pupil Size
95

Figure 5.16 Boxplot of standardized pupil size per visual sensation to illuminance intensity
(second round).
Table 5.6 One-way ANOVA test: Standardized Pupil Size versus Sensation
Source DF Adj SS Adj MS F-Value P-Value
Sensation 5 112.5 22.4962 2708.71 0.000
Error 22673 188.3 0.0083
Total 22678 300.8

S R-sq R-sq(adj) R-sq(pred)
0.0911324 37.40% 37.38% 37.36%

Means

Sensation N Mean StDev 95% CI
-2 970 0.08933 0.07080 ( 0.08359, 0.09506)
-1 2650 0.08595 0.11569 ( 0.08248, 0.08942)
0 5419 0.00821 0.09214 ( 0.00578, 0.01064)
1 5732 -0.04047 0.09528 (-0.04283, -0.03811)
2 4299 -0.07113 0.08346 (-0.07385, -0.06840)
3 3609 -0.14441 0.07450 (-0.14739, -0.14144)
3 2 1 0 -1 -2
30%
20%
10%
0%
-10%
-20%
-30%
Sensation
Individually Normalized Pupil Size (%)
Boxplot of Normalized Pupil Size
96

Pooled StDev = 0.0911324

For comparisons in different physiological categories, same procedure was adopted as done
in first round. The calculated pupil size change rate and other findings in different groupings will
be explained in Section 5.4.

Figure 5.17 Interval plot of comparisons of overall standardized pupil size per visual sensation
between eye color groups (second round).

Eye Color
Sensation_1
Brown Blue
3_Bright 2_Neutral 1_Dark 3_Bright 2_Neutral 1_Dark
15%
10%
5%
0%
-5%
-10%
Individually Normalized Pupil Size (%)
95% CI for the Mean
Interval Plot of Normalized Pupil Size
97

Figure 5.18 Boxplot of comparisons of overall standardized pupil size per visual sensation
between eye color groups (second round).
Table 5.7 One-way ANOVA test: Standardized Pupil Size versus Sensation between Eye Colors
ANOVA for Blue ANOVA for Brown
Analysis of Variance

Source DF Adj SS Adj MS F-Value
P-Value
Sensation_1 2 32.81 16.4074 1359.32
0.000
Error 5399 65.17 0.0121
Total 5401 97.98

Model Summary

S R-sq R-sq(adj) R-sq(pred)
0.109865 33.49% 33.47% 33.41%

Analysis of Variance

Source DF Adj SS Adj MS F-Value
P-Value
Sensation_1 2 57.48 28.7405 3421.08
0.000
Error 17274 145.12 0.0084
Total 17276 202.60

Model Summary

S R-sq R-sq(adj) R-sq(pred)
0.0916570 28.37% 28.36% 28.35%

Eye Color
Sensation_1
Brown Blue
3_Bright 2_Neutral 1_Dark 3_Bright 2_Neutral 1_Dark
50%
40%
30%
20%
10%
0%
-10%
-20%
-30%
-40%
Individually Normalized Pupil Size (%)
Boxplot of Normalized Pupil Size
98

Means

Sensation_1 N Mean StDev
95% CI
1_Dark 894 0.12850 0.14277
( 0.12130, 0.13570)
2_Neutral 1401 0.00422 0.09837
(-0.00153, 0.00997)
3_Bright 3107 -0.08329 0.10373
(-0.08715, -0.07942)

Pooled StDev = 0.109865

Means

Sensation_1 N Mean StDev
95% CI
1_Dark 2726 0.07320 0.08579
( 0.06976, 0.07664)
2_Neutral 4018 0.00960 0.08984
( 0.00677, 0.01244)
3_Bright 10533 -0.075968 0.093785
(-0.077719, -0.074218)

Pooled StDev = 0.0916570

Figure 5.19 Interval plot of comparisons of overall standardized pupil size per visual sensation
between age groups (second round).

Age Group
Sensation_1
Senior Junior
3_Bright 2_Neutral 1_Dark 3_Bright 2_Neutral 1_Dark
10%
5%
0%
-5%
-10%
Individually Normalized Pupil Size (%)
95% CI for the Mean
Interval Plot of Normalized Pupil Size
99

Figure 5.20 Boxplot of comparisons of overall standardized pupil size per visual sensation
between age groups (second round)
Table 5.8 One-way ANOVA test: Standardized Pupil Size versus Sensation between Age Groups
ANOVA for Junior ANOVA for Senior
Analysis of Variance

Source DF Adj SS Adj MS F-Value
P-Value
Sensation_1 2 32.84 16.4222 2231.37
0.000
Error 9054 66.63 0.0074
Total 9056 99.48

Model Summary

S R-sq R-sq(adj) R-sq(pred)
0.0857886 33.02% 33.00% 32.97%

Analysis of Variance

Source DF Adj SS Adj MS F-Value
P-Value
Sensation_1 2 55.47 27.7341 2598.44
0.000
Error 13619 145.36 0.0107
Total 13621 200.83

Model Summary

S R-sq R-sq(adj) R-sq(pred)
0.103312 27.62% 27.61% 27.58%

Means Means
Age Group
Sensation_1
Senior Junior
3_Bright 2_Neutral 1_Dark 3_Bright 2_Neutral 1_Dark
30%
20%
10%
0%
-10%
-20%
-30%
-40%
Individually Normalized Pupil Size (%)
Boxplot of Normalized Pupil Size
100

Sensation_1 N Mean StDev
95% CI
1_Dark 1677 0.07429 0.08151
( 0.07018, 0.07839)
2_Neutral 2651 0.00699 0.08186
( 0.00373, 0.01026)
3_Bright 4729 -0.07863 0.08935
(-0.08107, -0.07618)

Pooled StDev = 0.0857886

Sensation_1 N Mean StDev
95% CI
1_Dark 1943 0.09771 0.12155
( 0.09311, 0.10230)
2_Neutral 2768 0.00938 0.10102
( 0.00553, 0.01323)
3_Bright 8911 -0.07711 0.09963
(-0.07925, -0.07496)

Pooled StDev = 0.103312

Figure 5.21 Interval plot of comparisons of overall standardized pupil size per visual sensation
between myopic groups (second round).

Myopic
Sensation_1
Y N
3_Bright 2_Neutral 1_Dark 3_Bright 2_Neutral 1_Dark
10%
5%
0%
-5%
-10%
Individually Normalized Pupil Size (%)
95% CI for the Mean
Interval Plot of Normalized Pupil Size
101

Figure 5.22 Boxplot of comparisons of overall standardized pupil size per visual sensation
between myopic groups (second round)
Table 5.9 One-way ANOVA test: Standardized Pupil Size versus Sensation between Myopic
Groups
ANOVA for N ANOVA for Y
Analysis of Variance

Source DF Adj SS Adj MS F-Value
P-Value
Sensation_1 2 53.96 26.9776 2533.83
0.000
Error 14489 154.26 0.0106
Total 14491 208.22

Model Summary

S R-sq R-sq(adj) R-sq(pred)
0.103184 25.91% 25.90% 25.88%
Analysis of Variance

Source DF Adj SS Adj MS F-Value
P-Value
Sensation_1 2 35.09 17.5462 2616.56
0.000
Error 8184 54.88 0.0067
Total 8186 89.97

Model Summary

S R-sq R-sq(adj) R-sq(pred)
0.0818890 39.00% 38.99% 38.96%
Myopic
Sensation_1
Y N
3_Bright 2_Neutral 1_Dark 3_Bright 2_Neutral 1_Dark
30%
20%
10%
0%
-10%
-20%
-30%
Individually Normalized Pupil Size (%)
Boxplot of Normalized Pupil Size
102

Means

Sensation_1 N Mean StDev
95% CI
1_Dark 2311 0.09793 0.11547
( 0.09372, 0.10213)
2_Neutral 3518 0.00719 0.10053
( 0.00378, 0.01060)
3_Bright 8663 -0.06724 0.10075
(-0.06941, -0.06507)

Pooled StDev = 0.103184

Means

Sensation_1 N Mean StDev
95% CI
1_Dark 1309 0.06732 0.08171
( 0.06288, 0.07175)
2_Neutral 1901 0.01010 0.07416
( 0.00642, 0.01378)
3_Bright 4977 -0.09573 0.08470
(-0.09800, -0.09345)

Pooled StDev = 0.0818890

Figure 5.23 Interval plot of comparisons of overall standardized pupil size per visual sensation
between gender groups (second round).

Gender
Sensation_1
Male Female
3_Bright 2_Neutral 1_Dark 3_Bright 2_Neutral 1_Dark
10%
5%
0%
-5%
-10%
Individually Normalized Pupil Size (%)
95% CI for the Mean
Interval Plot of Normalized Pupil Size
103

Figure 5.24 Boxplot of comparisons of overall standardized pupil size per visual sensation
between gender groups (second round)
Table 5.10 One-way ANOVA test: Standardized Pupil Size versus Sensation between Gender
Groups
ANOVA for Female ANOVA for Male
Analysis of Variance

Source DF Adj SS Adj MS F-Value
P-Value
Sensation_1 2 27.75 13.8759 1253.97
0.000
Error 8758 96.91 0.0111
Total 8760 124.66

Model Summary

S R-sq R-sq(adj) R-sq(pred)
0.105193 22.26% 22.24% 22.20%
Analysis of Variance

Source DF Adj SS Adj MS F-Value
P-Value
Sensation_1 2 56.00 28.0006 3387.08
0.000
Error 13915 115.03 0.0083
Total 13917 171.03

Model Summary

S R-sq R-sq(adj) R-sq(pred)
0.0909223 32.74% 32.73% 32.71%
Gender
Sensation_1
Male Female
3_Bright 2_Neutral 1_Dark 3_Bright 2_Neutral 1_Dark
30%
20%
10%
0%
-10%
-20%
-30%
-40%
Individually Normalized Pupil Size (%)
Boxplot of Normalized Pupil Size
104

Means

Sensation_1 N Mean StDev
95% CI
1_Dark 828 0.08167 0.11801
( 0.07450, 0.08883)
2_Neutral 1805 0.00810 0.10854
( 0.00324, 0.01295)
3_Bright 6128 -0.08453 0.10232
(-0.08716, -0.08189)

Pooled StDev = 0.105193

Means

Sensation_1 N Mean StDev
95% CI
1_Dark 2792 0.08840 0.10153
( 0.08502, 0.09177)
2_Neutral 3614 0.00827 0.08277
( 0.00530, 0.01123)
3_Bright 7512 -0.07201 0.09050
(-0.07407, -0.06996)

Pooled StDev = 0.0909223

5.3 – Third Round
Similarly, the same analysis method and procedure was adopted for round three. Boxplots for the
purpose of illustrating quartile distributions of data and interval plots for the purpose of
illustrating estimated mean value with 95% confidence interval for whole data set, and each
categorized groupings were displayed.
105

Figure 5.25 Standardized pupil size distribution in each subject’s test (“No.” indicates a subject
ID).

No.
Sensation
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
3
2
1
0
-1
-2
-3
40%
30%
20%
10%
0%
-10%
-20%
-30%
Individually Normalized Pupil Size (%)
Boxplot of Normalized Pupil Size
106

Figure 5.26 Overall standardized pupil size distribution per visual sensation to illuminance
intensity (third round).

40% 20% 0% -20% -40%
-3
-2
-1
0
1
2
3
Individually Normalized Pupil Size (%)
Sensation
Each symbol represents up to 27 observations.
Dotplot of Normalized Pupil Size
107

Figure 5.27 Interval plot of standardized pupil size per visual sensation to illuminance intensity
(third round).
The pupil size patterns for visual sensations based on the combined data of all individuals are
shown in Figure 5.26. Overall, the standardized pupil sizes decreased while the generated
illuminance intensity was increasing. Figure 5.27 contains basically the same data as Figure 5.26,
but it shows a 95% confidence interval for pupil sizes per visual sensation. The interval lines are
clearly differentiated from each other among all sensation levels. Larger pupil size change was
observed in “dark” zones than in “bright” zones.

3 2 1 0 -1 -2 -3
15%
10%
5%
0%
-5%
-10%
Sensation
Individually Normalized Pupil Size (%)
95% CI for the Mean
Interval Plot of Normalized Pupil Size
108

Figure 5.28 Boxplot of standardized pupil size per visual sensation to illuminance intensity (third
round).
Table 5.11 One-way ANOVA test: Standardized Pupil Size versus Sensation
Analysis of Variance

Source DF Adj SS Adj MS F-Value P-Value
Sensation 6 76.18 12.6965 2310.34 0.000
Error 22453 123.39 0.0055
Total 22459 199.57

Model Summary

S R-sq R-sq(adj) R-sq(pred)
0.0741316 38.17% 38.16% 38.13%

Means

Sensation N Mean StDev 95% CI
-3 1063 0.12529 0.08196 ( 0.12084, 0.12975)
-2 2218 0.06768 0.06654 ( 0.06460, 0.07077)
3 2 1 0 -1 -2 -3
30%
20%
10%
0%
-10%
-20%
-30%
Sensation
Individually Normalized Pupil Size (%)
Boxplot of Normalized Pupil Size
109
-1 3807 0.04079 0.07623 ( 0.03844, 0.04315)
0 4156 0.000000 0.047034 (-0.002254, 0.002254)
1 5580 -0.03203 0.08503 ( -0.03397, -0.03008)
2 4448 -0.07849 0.08215 ( -0.08067, -0.07631)
3 1188 -0.09961 0.06347 ( -0.10383, -0.09539)

Pooled StDev = 0.0741316

For comparisons in different physiological categories, same procedure was adopted as done
in first round. The calculated pupil size change rate and other findings in different groupings will
be explained in Section 5.4.

Eye Color
Sensation_1
Brown Blue
3_Bright 2_Neutral 1_Dark 3_Bright 2_Neutral 1_Dark
8%
6%
4%
2%
0%
-2%
-4%
-6%
-8%
Individually Normalized Pupil Size (%)
95% CI for the Mean
Interval Plot of Normalized Pupil Size
110
Figure 5.29 Interval plot of comparisons of overall standardized pupil size per visual sensation
between eye color groups (third round).

Figure 5.30 Boxplot of comparisons of overall standardized pupil size per visual sensation
between eye color groups (third round).
Table 5.12 One-way ANOVA test: Standardized Pupil Size versus Sensation between Eye
Colors
ANOVA for Blue ANOVA for Brown
Analysis of Variance

Source DF Adj SS Adj MS F-Value
Analysis of Variance

Source DF Adj SS Adj MS F-Value
Eye Color
Sensation_1
Brown Blue
3_Bright 2_Neutral 1_Dark 3_Bright 2_Neutral 1_Dark
30%
20%
10%
0%
-10%
-20%
-30%
Individually Normalized Pupil Size (%)
Boxplot of Normalized Pupil Size
111
P-Value
Sensation_1 2 15.20 7.60241 1143.19
0.000
Error 7616 50.65 0.00665
Total 7618 65.85

Model Summary

S R-sq R-sq(adj) R-sq(pred)
0.0815486 23.09% 23.07% 23.04%

P-Value
Sensation_1 2 47.87 23.9372 4354.97
0.000
Error 14838 81.56 0.0055
Total 14840 129.43

Model Summary

S R-sq R-sq(adj) R-sq(pred)
0.0741385 36.99% 36.98% 36.97%

Means

Sensation_1 N Mean StDev
95% CI
1_Dark 2438 0.07320 0.08189
( 0.06996, 0.07644)
2_Neutral 1606 0.00000 0.04931
(-0.00399, 0.00399)
3_Bright 3575 -0.02853 0.09223
(-0.03120, -0.02586)

Pooled StDev = 0.0815486

Means

Sensation_1 N Mean StDev
95% CI
1_Dark 4650 0.05595 0.07801
( 0.05381, 0.05808)
2_Neutral 2550 0.000000 0.045551
(-0.002878, 0.002878)
3_Bright 7641 -0.071219 0.079244
(-0.072881, -0.069556)

Pooled StDev = 0.0741385

Age Group
Sensation_1
Senior Junior
3_Bright 2_Neutral 1_Dark 3_Bright 2_Neutral 1_Dark
8%
6%
4%
2%
0%
-2%
-4%
-6%
-8%
Individually Normalized Pupil Size (%)
95% CI for the Mean
Interval Plot of Normalized Pupil Size
112
Figure 5.31 Interval plot of comparisons of overall standardized pupil size per visual sensation
between age groups (third round).

Figure 5.32 Boxplot of comparisons of overall standardized pupil size per visual sensation
between age groups (third round)
Table 5.13 One-way ANOVA test: Standardized Pupil Size versus Sensation between Age
Groups
ANOVA for Senior ANOVA for Junior
Analysis of Variance

Source DF Adj SS Adj MS F-Value
Analysis of Variance

Source DF Adj SS Adj MS F-Value
Age Group
Sensation_1
Senior Junior
3_Bright 2_Neutral 1_Dark 3_Bright 2_Neutral 1_Dark
30%
20%
10%
0%
-10%
-20%
-30%
Individually Normalized Pupil Size (%)
Boxplot of Normalized Pupil Size
113
P-Value
Sensation_1 2 36.16 18.0823 2729.78
0.000
Error 13453 89.11 0.0066
Total 13455 125.28

Model Summary

S R-sq R-sq(adj) R-sq(pred)
0.0813884 28.87% 28.86% 28.84%

P-Value
Sensation_1 2 22.73 11.3673 2272.26
0.000
Error 9001 45.03 0.0050
Total 9003 67.76

Model Summary

S R-sq R-sq(adj) R-sq(pred)
0.0707293 33.55% 33.54% 33.51%

Means

Sensation_1 N Mean StDev
95% CI
1_Dark 3757 0.05082 0.07795
( 0.04822, 0.05342)
2_Neutral 2192 0.000000 0.045067
(-0.003407, 0.003407)
3_Bright 7507 -0.06700 0.09077
( -0.06884, -0.06516)

Pooled StDev = 0.0813884

Means

Sensation_1 N Mean StDev
95% CI
1_Dark 3331 0.07435 0.07999
( 0.07195, 0.07675)
2_Neutral 1964 0.00000 0.04915
(-0.00313, 0.00313)
3_Bright 3709 -0.03862 0.07154
(-0.04089, -0.03634)

Pooled StDev = 0.0707293

114

Figure 5.33 Interval plot of comparisons of overall standardized pupil size per visual sensation
between myopic groups (third round).

Myopic
Sensation_1
Y N
3_Bright 2_Neutral 1_Dark 3_Bright 2_Neutral 1_Dark
8%
6%
4%
2%
0%
-2%
-4%
-6%
-8%
Individually Normalized Pupil Size (%)
95% CI for the Mean
Interval Plot of Normalized Pupil Size
115

Figure 5.34 Boxplot of comparisons of overall standardized pupil size per visual sensation
between myopic groups (third round)
Table 5.14 One-way ANOVA test: Standardized Pupil Size versus Sensation between Myopic
Groups
ANOVA for N ANOVA for Y
Analysis of Variance

Source DF Adj SS Adj MS F-Value
P-Value
Sensation_1 2 46.44 23.2207 3602.80
0.000
Error 13374 86.20 0.0064
Total 13376 132.64

Model Summary

S R-sq R-sq(adj) R-sq(pred)
0.0802820 35.01% 35.00% 34.99%
Analysis of Variance

Source DF Adj SS Adj MS F-Value
P-Value
Sensation_1 2 17.15 8.57710 1585.26
0.000
Error 9080 49.13 0.00541
Total 9082 66.28

Model Summary

S R-sq R-sq(adj) R-sq(pred)
0.0735564 25.88% 25.86% 25.84%
Myopic
Sensation_1
Y N
3_Bright 2_Neutral 1_Dark 3_Bright 2_Neutral 1_Dark
30%
20%
10%
0%
-10%
-20%
-30%
Individually Normalized Pupil Size (%)
Boxplot of Normalized Pupil Size
116

Means

Sensation_1 N Mean StDev
95% CI
1_Dark 4080 0.06706 0.08533
( 0.06460, 0.06952)
2_Neutral 2487 0.000000 0.046864
(-0.003155, 0.003155)
3_Bright 6810 -0.06702 0.08658
( -0.06893, -0.06512)

Pooled StDev = 0.0802820

Means

Sensation_1 N Mean StDev
95% CI
1_Dark 3008 0.05485 0.07098
( 0.05222, 0.05748)
2_Neutral 1669 -0.00000 0.04730
(-0.00353, 0.00353)
3_Bright 4406 -0.04306 0.08286
(-0.04524, -0.04089)

Pooled StDev = 0.0735564

Figure 5.35 Interval plot of comparisons of overall standardized pupil size per visual sensation
between gender groups (third round).

Gender
Sensation_1
Male Female
3_Bright 2_Neutral 1_Dark 3_Bright 2_Neutral 1_Dark
8%
5%
3%
0%
-3%
-5%
Individually Normalized Pupil Size (%)
95% CI for the Mean
Interval Plot of Normalized Pupil Size
117

Figure 5.36 Boxplot of comparisons of overall standardized pupil size per visual sensation
between gender groups (third round)
Table 5.15 One-way ANOVA test: Standardized Pupil Size versus Sensation between Gender
Groups
ANOVA for Female ANOVA for Male
Analysis of Variance

Source DF Adj SS Adj MS F-Value
P-Value
Sensation_1 2 20.51 10.2548 1409.30
0.000
Error 9753 70.97 0.0073
Total 9755 91.48

Model Summary

S R-sq R-sq(adj) R-sq(pred)
0.0853028 22.42% 22.40% 22.38%
Analysis of Variance

Source DF Adj SS Adj MS F-Value
P-Value
Sensation_1 2 41.39 20.6967 3999.77
0.000
Error 12701 65.72 0.0052
Total 12703 107.11

Model Summary

S R-sq R-sq(adj) R-sq(pred)
0.0719337 38.64% 38.63% 38.62%
Gender
Sensation_1
Male Female
3_Bright 2_Neutral 1_Dark 3_Bright 2_Neutral 1_Dark
30%
20%
10%
0%
-10%
-20%
-30%
Individually Normalized Pupil Size (%)
Boxplot of Normalized Pupil Size
118

Means

Sensation_1 N Mean StDev
95% CI
1_Dark 2411 0.05607 0.08461
( 0.05266, 0.05947)
2_Neutral 1650 0.00000 0.04693
(-0.00412, 0.00412)
3_Bright 5695 -0.05246 0.09379
(-0.05468, -0.05025)

Pooled StDev = 0.0853028

Means

Sensation_1 N Mean StDev
95% CI
1_Dark 4677 0.06488 0.07701
( 0.06281, 0.06694)
2_Neutral 2506 -0.000000 0.047110
(-0.002817, 0.002817)
3_Bright 5521 -0.06292 0.07665
( -0.06482, -0.06103)

Pooled StDev = 0.0719337

5.4 – Discussions between different rounds
Sections 5.1, 5.2 and 5.3 mainly discussed pupil size change per visual sensation under different
lighting settings as well as comparisons between categorized groupings of human subjects
according to physiological features. Checking consistency of previous comparisons in different
experiment rounds and discuss other potential findings, especially difference between
experiment rounds, is also very important.
5.4.1- Checking Consistency of Previous Observations
By comparing overall standardized pupil size distribution of first round (Figure 5.2, Figure 5.3),
second round (Figure 5.14, Figure 5.15) and third round (Figure 5.26, Figure 5.27), the pupil size
patterns for visual sensations based on the combined data of all individuals follow similar trend
among different rounds, which is, the standardized pupil sizes decreased while the generated
illuminance intensity was increasing. And the mean value per sensation is mostly clearly
differentiated from each other. Furthermore, the interval of each sensation is relatively small
which indicates the potential use as baseline for each sensation under those lighting conditions.
By grouping human subjects into different categories based on their physiological features, a
better understanding of those physiological characters could be achieved. Through comparisons
119
between eye colors among all three rounds (Figure 5.5, Figure 5.18 and Figure 5.31), it can be
concluded that the eye color has effect on pupil size change. Subjects with brown eye color had
less pupil size change when reported dark sensations. While, subjects with blue eye color has
consistent large pupil size change in dark sensation than in bright sensation.
In comparisons between age groups among all three rounds (Figure 5.7, Figure 5.20 and
Figure 5.33), a consistent result was not found. There is larger pupil size change in the junior
group in the first round, larger pupil size change in the senior group in the second group and
similar pupil size pattern but different distribution in the third round. Therefore, age doesn’t
seem to have effect on pupil size change. A possible reason for this observation could be
restrictions on the samples or the standard of division to the group. 25 may not be the right
choice for division. And there were less samples in >25 group. A larger range of age should be
achieved and would be more precise.
Through comparisons between myopic and non-myopic groups among all three rounds
(Figure 5.9, Figure 5.22 and Figure 5.35), consistent effect of myopia on pupil size has been
observed. Myopic subjects have less pupil size change than the non-myopic group, especially
when in dark environment. It can be easily understood because myopic subjects have less ability
in controlling their pupils.
In the end, comparisons between gender groups among all three rounds (Figure 5.11, Figure
5.24 and Figure 5.37) don’t indicate any significant effect of gender on pupil size. This is
consistent with background research mentioned in Chapter 2.

120
5.4.2- Findings Observed from Comparisons between Experiment Rounds
The settings of three rounds differentiated in light color temperature and task types. It is also
important to discuss whether color temperature and task types have any effect on pupil size.
Comparisons between first round and second round, second round and third round were done
separately to test the effect of color temperature (Figure 5.37) and task type (Figure 5.38).

Figure 5.37 Interval plot of comparisons of overall standardized pupil size per visual sensation
between color temperatures.

Color
Temperature
Sensation_1
Warm Daylight
3_Bright 2_Neutral 1_Dark 3_Bright 2_Neutral 1_Dark
10%
5%
0%
-‐5%
-‐10%
Normalized
Pupil
Size
Interval
Plot
of
Normalized
Pupil
Size
95%
CI
for
the
Mean
Individual
standard
deviations
were
used
to
calculate
the
intervals.
121

Figure 5.38 Interval plot of comparisons of overall standardized pupil size per visual sensation
between task types.
Daylight seems to cause larger pupil size change than warm color temperature. A
computer-based task seems to cause a larger pupil size change than a paper-based task as well.

5.5 – Summary
After doing comparisons between different rounds, it is clear that categorized human subjects
share similar pupil size responses to different lighting conditions. There are also some particular
features in each group as well. Then, a summary of each round is shown in column figures for a
better graphical display (Figure 5.39, Figure 5.40 and Figure 5.41).
Task
Type
Sensation_1
Paper Computer
3_Bright 2_Neutral 1_Dark 3_Bright 2_Neutral 1_Dark
10%
5%
0%
-‐5%
-‐10%
Normalized
Pupil
Size
Interval
Plot
of
Normalized
Pupil
Size
95%
CI
for
the
Mean
Individual
standard
deviations
were
used
to
calculate
the
intervals.
122

Figure 5.39 Summary of average pupil size change in each category of first round.

Figure 5.40 Summary of average pupil size change in each category of second round.

6.9%
7.0%

8.2%

6.5%

14.1%

5.3%

2.2%

12.0%

7.8%

5.0%
5.1%

6.5%

4.8%

6.0%

4.8%

5.0%

5.2%
5.3%

0.0%

2.0%

4.0%

6.0%

8.0%

10.0%

12.0%

14.0%

16.0%

Male
Female
<25
>=25
Blue
Brown
Yes
No
Average

Gender
Age
Eye
color
Myopic
calculaJon

First
Round
(Computer
+
Warm)
Dark
Vs.
Neutral
Bright
Vs.
Neutral

9.1%

8.2%

7.6%

10.2%

13.0%

7.8%

7.1%

9.8%

9.1%

7.3%

8.0%

7.5%
7.4%

8.5%

8.2%

9.5%

7.0%

7.9%

0.0%

2.0%

4.0%

6.0%

8.0%

10.0%

12.0%

14.0%

16.0%

Male
Female
<25
>=25
Blue
Brown
Yes
No
Average

Gender
Age
Eye
color
Myopic
calculaJon

Second
Round
(Computer
+
Daylight)
Dark
Vs.
Neutral
Bright
Vs.
Neautral

123

Figure 5.41 Summary of average pupil size change in each category of third round.
With collected pupil size, pupil size calculations of human subjects groups based on
physiological divisions under different lighting conditions were conducted. All calculated results
could be used as reference in future lighting control.

6.4%

5.5%

7.5%

5.3%

7.5%

5.8%
5.7%

6.8%

6.3%

6.5%

5.1%

4.1%

6.8%

3.1%

7.3%

4.6%

6.8%

5.5%

0.0%

2.0%

4.0%

6.0%

8.0%

10.0%

12.0%

14.0%

16.0%

Male
Female
<25
>=25
Blue
Brown
Yes
No
Average

Gender
Age
Eye
color
Myopic
calculaJon

Third
Round
(Paper
+
Daylight)
Dark
Vs.
Neutral
Bright
Vs.
Neautral

124
Chapter 6: Conclusions of Study
The study was aiming at figuring out the potential use of pupil size for future office lighting
control. For the purpose of testing the relationship between pupil size and lighting condition and
excavating other potential effects, conducting human subject experiment was chosen as method
for the study. USC School of Architecture provided a chamber, and experimental devices were
installed based on the experiment design and requirements.
Three rounds of experiment were finished, and 20 volunteers participated in each round.
There was only one participant per experiment to be tested. Each experiment lasted on an
average of 1 hour and 20 minutes. During the experiment, participant was required to finish a
certain type of work (computer-based or paper-based) under different lighting conditions, and
visual responses to the lighting conditions were collected through questionnaire and pupil size.
Lighting parametric data were also monitored and stored automatically during the tests.
The relationship between pupil size and lighting condition was subjected to statistical
analysis. Differences among physiological groups of human subjects were also summarized as
important findings.
6.1 – Illuminance, Sensations and Pupil Size
Illuminance, sensation, and pupil size were selected as three significantly important parameters
consequently representing lighting condition, subjective feeling/judgment ,and objective
physiological response. Ten levels of illuminance ranging from 50 lux to 1400 lux with 150 lux
interval were created for each human subject involved experiment as an indicator. A 7-point
scale questionnaire was adopted to collect participant’s sensation per illuminance level. Through
the entire experiment, the pupilometer worn by the participant was monitoring and collecting
125
pupil size parametric data. By combining all those data, relationships among illuminance,
sensations, and pupil size could be analyzed and concluded. The potential use of pupil size was
also validated.
Preferred illuminance range differentiates significantly per visual sensation, and for each
sensation there is a corresponding illuminance range. Since average luminance and illuminance
have a linear regression relation with each other, for each sensation, there is also a corresponding
average luminance range. Sensations could reflect participants’ preference both objectively and
subjectively.
Standardized pupil size distribution has significant difference per visual sensation, which is
consistent through all three rounds of experiment. There is at least about 4% difference in pupil
size between visual sensations. It is certain that pupil size can be considered as an indicator of
preferred lighting conditions, which demonstrates the potential use of pupil size for lighting
control.
6.2 – Further Conclusions based on Physiological Features of Human Subjects
Based on the previous study and research purpose, age, gender, ethnicity, eye color and myopic
or non-myopic were chosen as physiological features and were collected from participants’
questionnaires. Restrictions had also be adopted to reduce effect on human subjects’ behaviors
due to other factors, which included food and drink prohibition during experiment, etc.
There were two groups in each category for the further analysis. The study was also trying to
balance number of people in each group to minimize the effect due to large difference in sample
size of each group in the category. Finally, among all three rounds of experiment, except eye
color category, all other categories have achieved 40%/60% distribution between two groups
126
which mostly satisfied study expectations. By applying classifications to pupil size data of
human subjects, more findings were concluded:
Comparisons of overall standardized pupil size per visual sensation between eye colors, age
groups, myopic groups as well as gender groups, present the differentiated standardized pupil
size distribution due to the physiological features of human subjects.
Eye color and myopia have a significant effect on pupil size change. Subjects with brown
eye color have an average of 4.2% less pupil size change compared with subjects with blue eye
color; myopic group has an average of 4.5% less pupil size than non-myopic group. However,
age and gender don’t indicate any consistent effect on pupil size, since average pupil size
difference between two age groups in three rounds are 3.4%, -2.5% and -0.5%; average pupil
size difference between gender groups in there rounds are very limited, which are -0.2%, 0.2%
and 2.3%. A previous study (Winn et al. 1994) indicated that gender doesn’t have effect on pupil
size. It is validated in this study.
6.3 – Color Temperature and Task Type
As mentioned in Chapter 3 and Chapter 4, color temperature and task type are two other
important factors studied in this research. Therefore, three rounds of experiment were conducted
separately with different lighting fixtures and other device installed: warm color temperature +
computer-based task, daylight color temperature + computer-based task, and daylight color
temperature + paper-based task. By conducting comparisons between these three groups, effect
of color temperature and task type was observed:
About 2.5% larger pupil size change has been observed in daylight group than in warm
group, and 2.7% larger in computer-based task group than in paper-based group.
127
6.4 – Potential Use of Findings
With the findings of pupil size, automatic lighting control based on pupil size may be possible.
Current devices, such as a pupilometer, smartphone or even Google glass (Google glass is a
smart-control device developed by Google, which is like goggle with very tiny glass for display
and a highly integrated projector installed, by using Google glass, it is very easy to take pictures
searching for information, just like using a smartphone) , make it possible for convenient and
in-time tracking and monitoring. When a tracking device identifies pupil size, judgment based on
the human physiological features and collected dataset will analyze and calculate out if current
lighting condition is appropriate, too light, or two dark for the specific subject. Then, a control
signal indicating the optimal amount of change to the lighting device will be sent. As a result,
lighting is adjusted. The conceptual controlling strategy is shown in Figure 6.1.

Figure 6.1 Conceptual strategy for automatic lighting controlling
This strategy could be adopted for both individual lighting control and group lighting control.
Or even, it provides idea for office lighting, layout design. By adopting this strategy, it is
Pupil
Size

Tracking

Judgment:

Human
physiological

features
&
collected

dataset
(%)

Adjust

Lighting

Device

128
believed that large amount of electricity could be saved for office lighting, then a more efficient,
environmental friendly working environment could be achieved.
129
Chapter 7: Future Work
Human factors, lighting parameters, and hardware and software combination could be improved.
There are also limitations when conducting this research, which could be paid more attention to
or solved in the future.
7.1 – Possible Improvements on Participants
Although the sample size (20 participants in each round) in this research is believed enough for
an efficient analysis, a larger sample size is always better to accessing more precise and stable
results for a human subject involved research. The bigger sample size could result in a more
reliable statistical result with narrower confidence interval range, which is more useful for the
future automatic control logic generation for the lighting.
Meanwhile, a better variety of participants should be also achieved and balanced in the
experiment. Each physiological featured group is best to be divided equally to reduce the
potential negative effect due to the different sample size in the compared groups. As the study
was conducted on the campus, most participants involved in the experiment are graduate and
undergraduate students which similar age. A better age group division should include more
middle-aged people and old people as well. More people with blue eye color should be recruited
to balance the eye color category. A detailed consideration on myopia should also be studied as
different glasses degree could also behave differently.
7.2 – Possible Improvements on Lighting Parameters
Illuminance was used as a major parameter representing the lighting condition. It also served as a
reference when creating different lighting conditions. Although luminance was calculated and
130
discussed as well, a deeper research on luminance and other lighting parameters should be
conducted in the future.
The relationship between illuminance and average luminance was discussed and a regression
was generated based on the collected. However, there is a range of luminance for the workstation
at each illuminance level. And distribution of luminance in the captured view from fisheye lance
varies a lot. In this case, not only average luminance value should be considered, the minimum,
maximum and distribution of luminance in the view should also be analyzed to reach a deeper
understanding of luminance effect on pupil size and visual sensations. In addition, other
parameters such as contrast ratio, light spectrum may also be included in the future study.
Daylight could also be introduced into this study since another major component in office
lighting design is combining daylight with artificial lighting. In this way, a better and more
efficient office lighting design could be achieved in the future by integrating automatic control
for both artificial lighting and daylight. This strategy will also have an effect on the façade
design and office layout.
7.3 – Possible Improvements on Hardware and Software
A lot investment has been put into this study for sensors, device, and software. There are more
than 40 LED lamps, there are two computers for monitoring and conducting task purposes. And
there is precious pupilometer tracking pupil size. But it still could be improved.
In the beginning of this study, there were blinking problems of light source due to the
unstable power supply in the chamber, especially when it was at low illuminance levels, which
caused negative effect on participant’s response to the satisfaction level. Although the problem
was minimized a lot by replacing the manual dimmer, it still has occasional blinking when other
131
big machine on the basement level, such as elevator, was in use. THe power supply could be
fixed or a better dimmer or self-designed and made dimmer could be used to fully satisfy
experiment requirement.
Sometimes, there were complains from participants about the glare problems that caused by
the reflective glass on the pupilometer from light source. A temporary solution, which used a cap
or a small paperboard to cover top of pupilometer, was introduced. But it blocked some view
above the pupilometer, which is not serious but unfavorable in the experiment. It might be
problem of the location of LED lamps, the distance of LED lamps to participant’s head is
relatively smaller than traditional office. Therefore, a better design of the light source should be
considered.
The light source used in the chamber was LED lamps, which are different from lighting
fixtures used in the current offices. Replacing LED lamps could not only achieve a better
office-like environment but also may solve glare problems.
Currently, there are too many devices on the table, which may be some distractions for the
participant. Two laptops were used and one was for pupilometer and the other was for
illuminance meters. And each laptop has software installed for collecting data purpose. If those
two data collection programs could be combined, there would be easier to monitor and collection.
The combined data would also be more sufficient since it saved a large amount of time for
preliminary data packaging and preparation. The combined program only requires one computer
and saves space in the chamber as a result. In addition, it could also integrate dimming control in
the program so as to be more accessible to different lighting conditions.
132
7.4 – Strategy Development
A calculated pupil size change based on different conditions was summarized in this study. But
detailed control strategy and formula based on the collected database should be worked in the
future. Some validation research should be conducted for the strategy to ensure its proficiency.
With a detailed control strategy and formula, software could be programmed and developed
for controlling lighting devices or even other self-designed physical applications. However, real
testing of all those software and physical applications is very necessary. Commissioning in a real
project would be the most valuable. Both short term and long term periods commissioning are
highly recommended.
As mentioned in the Chapter 6, the strategy is not restricted to the office lighting control. It
could also be developed for other purposes but with further related research. Pupil size based
control strategy is believed to have broader use in the industry. More related research are highly
recommended and valued.

133
Bibliography
“6 Tips for Taking Better Macro Photos with the iPhone Camera.” 2012.
http://osxdaily.com/2012/10/07/tips-better-iphone-macro-photos/.

Berman, B Y S A M. 2000. “The Coming Revolution in Lighting Practice.”

Campbell, B Y F W, and D G Green. 1965. “Optical and Retinal Factors Affecting Visual
Resolution.” J.Physiol 181: 576–93.

Campbell, B Y F W, and R W Gubisch. 1966. “Optical Quality of The Human Eye.” J.Physiol
186: 558–78.

Campbell, F.W. 1957. “The Depth of Field of the Human Eye.” Optica Acta 4: 157–64.
Deneen, S. 2004. Seeing the Light.

Dilouie, Craig. 2003. “Lighting and Productivity : Missing Link Found ?” Architectural Lighting
18,6: 39.

Eastman, A.A., and J.F. McNelis. 1963. “An Evaluation of Sodium, Mercury and Filament
Lighting for Roadways.” Illum Eng.

“Event Horizon Volume 3 6 Aging Eyes and Pupil Size.” 2014. Accessed June 22.
http://amateurastronomy.org/Events/EH361.html.

Ferguson, H.M., and W.R. Stevens. 1956. “Relative Brightness of Colored Light Sources.” Trans
IES.

Hedge, Alan, William R Sims Jr., and Franklin D Becker. 1990. “CUergo: Cornell University
Lighting Research Study.” http://ergo.human.cornell.edu/lighting/lilstudy/lilstudy.htm.

Irlen, Helen. 1991. “Reading By The Colors Overcoming Dyslexia and Other Reading
Disabilities Through the Irlen Method” 27 (Chapter 3).

Leibowitz, H. 1952. “The Effect of Pupil Size on Visual Acuity for Photometrically Equated
Test Fields at Various Levels of Luminance.” Journal of the Optical Society of America 42
(6): 416–22. http://www.ncbi.nlm.nih.gov/pubmed/14939110.

“Light+Architecture: Lighting Simulation Software: Autodesk 3DS MAX.” 2011.
http://blog.lightingvanguard.com/2011/08/lighting-simulation-software-autodesk.html.

“Lighting Analysis in BIM | Sustainability Workshop.” 2014. Accessed June 22.
http://sustainabilityworkshop.autodesk.com/buildings/lighting-analysis-bim.

Luckiesh, M., and F.K. Moss. 1934. “Seeing in Sodium Vapor Light.” J Opt Soc Am.

134
Rapp, D. 1997. Is This Your Child’s World? New York.

Winn, B, D Whitaker, D B Elliott, and N J Phillips. 1994. “Factors Affecting Light-Adapted
Pupil Size in Normal Human Subjects.” Investigative Ophthalmology & Visual Science 35
(3): 1132–37. http://www.ncbi.nlm.nih.gov/pubmed/8125724.

Cassin, B., & Solomon, S. (1990). Dictionary of Eye Terminology. Gainsville, Florida: Triad
Publishing Company.
(2010). In W. T. Grondzik, A. G. Kwok, B. Stein, & J. S. Reynolds, Mechanical and Electrical
Equipment for Buildings. John Wiley & Sons, Inc.
Head Mounted Eye Tracking See the World Through Different Eyes. (n.d.). 2014
http://www.mangold-international.com/eye-tracking/head-mounted/overview.html
LabVIEW System Design Software. (n.d.) 2014. http://www.ni.com/labview/
OMNITAB 80. (n.d.) 2014. http://www.nist.gov/itl/sed/omnitab-80.cfm
Choi, Joon-Ho, Vivian Loftness, and Azizan Aziz. 2012. “Post-Occupancy Evaluation of 20
Office Buildings as Basis for Future IEQ Standards and Guidelines.” Energy and Buildings
46 (March): 167–175. doi:10.1016/j.enbuild.2011.08.009.
http://linkinghub.elsevier.com/retrieve/pii/S0378778811003434.

Manav, Banu. 2007. “An Experimental Study on the Appraisal of the Visual Environment at
Offices in Relation to Colour Temperature and Illuminance.” Building and Environment 42
(2) (February): 979–983. doi:10.1016/j.buildenv.2005.10.022.
http://linkinghub.elsevier.com/retrieve/pii/S036013230500452X.

Oi, Naoyuki, and Hironobu Takahashi. “PREFERRED COMBINATIONS BETWEEN
ILLUMINANCE AND COLOR TEMPERATURE IN SEVERAL SETTINGS FOR
DAILY LIVING ACTIVITIES”: 2–5.

Sauro, Jeff. 2014. “Should You Use 5 or 7 Point Scales?” Accessed March 12.
https://www.measuringusability.com/blog/scale-points.php.

Sivaji, Ashok, Sajidah Shopian, Zulkifle Mohd Nor, Ngip-Khean Chuan, and Shamsul Bahri.
2013. “Lighting Does Matter: Preliminary Assessment on Office Workers.” Procedia -
Social and Behavioral Sciences 97 (November): 638–647.
doi:10.1016/j.sbspro.2013.10.283.
http://linkinghub.elsevier.com/retrieve/pii/S1877042813037282.
135

Veitch, J.A., and G.R. Newsham. 1998. “Lighting Quality and Energy-Efficiency Effects on
Task Performance, Mood, Health, Satisfaction, and Comfort.” Journal of the Illuminating
Engineering Society 27 (1) (January): 107–129. doi:10.1080/00994480.1998.10748216.
http://dx.doi.org/10.1080/00994480.1998.10748216.

Wright, K P, P Badia, B L Myers, S C Plenzler, and M Hakel. 1997. “Caffeine and Light Effects
on Nighttime Melatonin and Temperature Levels in Sleep-Deprived Humans.” Brain
Research 747 (1) (January 30): 78–84. http://www.ncbi.nlm.nih.gov/pubmed/9042530.

Abstract (if available)

Abstract The goal of this research is to establish a visual environment diagnostic model based on the occupant’s physiological responses for detecting improper ambient lighting conditions, a major contributing factor to visual stress and work productivity in office workplace environments. The human body, as a biological mechanism, naturally minimizes the effects of ambient environmental stressors using its physiological autonomous nerve system. This system enables a human’s pupils to dilate and contract, depending on visual sensations affected by the ambient lighting conditions. An extensive experiment using human subjects will be conducted in an environmental chamber on the University of Southern California campus. All parametric data including human pupil sizes and lighting parameters will be categorized by age and ethnic origin, to investigate and determine the most common features of pupil sizes per visual sensation among individuals. Lighting parameters, including illuminance (lux), luminance (cd/m2), and lighting‐color‐temperature (K), will be controlled and maintained for each volunteer subject based on his/her task‐type (computer‐based or paper‐based), which is most typical in the current office environment. ❧ This study will provide unique knowledge concerning how an occupant via his/her physiological signal, i.e. pupil size can interact with the visual (lighting) environment. The research outcome will be potentially applicable in reality to diagnose the lighting quality in workplace environments, and to integrate an occupant’s pupil size information for the visual environmental controls.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Human–building integration: machine learning–based and occupant eye pupil size–driven lighting control as an applicable visual comfort tool in the office environment

PDF

Human-building integration: Investigation of human eye pupil sizes as a measure of visual sensation in the workstation environment

PDF

Impacts of indoor environmental quality on occupants environmental comfort: a post occupancy evaluation study

PDF

Enhancing thermal comfort: air temperature control based on human facial skin temperature

PDF

Indoor environmental quality and comfort: IEQ adaptation and human physiological responses in commercial buildings

PDF

Developing a data-driven model of overall thermal sensation based on the use of human physiological information in a built environment

PDF

Developing environmental controls using a data-driven approach for enhancing environmental comfort and energy performance

PDF

Multi-occupancy environmental control for smart connected communities

PDF

IEQ, sleep quality, and IoT: meta-analysis on improving IEQ and sleep quality using IoT

PDF

Indoor air quality for human health in residential buildings

PDF

Impacts of building performance on occupants' work productivity: a post occupancy evaluation study

PDF

Adaptive façade controls: A methodology based on occupant visual comfort preferences and cluster analysis

PDF

Enhancing thermal comfort: data-driven approach to control air temperature based on facial skin temperature

PDF

Energy savings by using dynamic environmental controls in the cavity of double skin facades

PDF

Exploration for the prediction of thermal comfort & sensation with application of building HVAC automation

PDF

Considering occupants: comprehensive POE research on office environment of Southern California

PDF

Enhanced post occupancy evaluation (POE) for office building: improvement of current methodology to identify impact of ambient environment

PDF

Double skin façades performance: effects on daylight and visual comfort in office spaces

PDF

Data visualization in VR/AR: static data analysis in buildings

PDF

Quantify human experience: integrating virtual reality, biometric sensors, and machine learning

Asset Metadata

Creator Zhu, Rui (author)

Core Title Human-environmental interaction: potential use of pupil size for office lighting controls

School School of Architecture

Degree Master of Building Science

Degree Program Building Science

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag human comfort,indoor environmental quality,lighting control,OAI-PMH Harvest,office,pupil size,sustainability

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Choi, Joon-Ho (committee chair), Kensek, Karen M. (committee member), Noble, Douglas (committee member)

Creator Email ruizhu@usc.edu,zhur08thu@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-464517

Unique identifier UC11287927

Identifier etd-ZhuRui-2847.pdf (filename),usctheses-c3-464517 (legacy record id)

Legacy Identifier etd-ZhuRui-2847.pdf

Dmrecord 464517

Document Type Thesis

Rights Zhu, Rui

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA