Study Effects that Factors Have On Human`s Vision Capability

1
Study Effects that Factors Have On Humans
Vision Capability

• DOE

JI Xiaocong Huang Fei Wang Mingqiang Mei Lin
2
Introduction

• DOE

3
We are thinking about
Under what condition can he see it clearly?
4
Objectives of the experiment

• Measure the effects of aimed environmental
  factors on humans vision
• luminance, color, shape and contrast

• preliminary work
• Indentify the response factor vision ability
• Indentify the degrees to which these factors
  contribute
• Reference
• Based on National standard vision chart - the
  standard logarithmic visual acuity chart GB
  11533-1989

5
Variables(1/3)

• Control variables
• Screen luminance
• Character color Screen background color
• Character shape
• Screen contrast
• Character shape

6
Variables(2/3)

• "held constant" variables
• Subjects age
• Subjects vision
• Subjects health condition
• Subjects environmental adaptability
• Subjects visual angle
• Illumination of the test environment
• Testing time
• The interval between vision table and subjects

7
Variables(3/3)

• Nuisance factor (units)
• Gene
• Gender
• Psychological state

8
Restrictions

• Focus on the environmental factors that affect
  the vision capability Ignore the individual
  differences between anticipants
• Restricted by the experiment conditions, some of
  the control variables are difficult to adjust.

9
Methodology

• DOE

10
Analysis of the factors(1/2)
Two Level
11
Analysis of the factors(2/2)
12
Analysis of the Response

• Based on National standard vision chart - the
  standard logarithmic visual acuity chart ?GB
  11533-1989?
• The ability to recognize the different layout (12
  degrees )

13
Full Factorial VS Fractional Factorial

• Full factorial design
• 5 factorials
• 2532

• Full factorial design
• 4 factorials
• 2416

• Not easy to actualize
• Lead the tiredness

• All the factors is important

Which one is better??

• Fractional Factorial
• 5 factorials
• 2(5-1)16

14
Fractional Factorial Design
Alias Structure I ABCDE Blocks A
BCDE B ACDE C ABDE D ABCE E
ABCD
Blocks(Cont) AB CDE AC BDE AD
BCE AE BCD BC ADE BD ACE BE
ACD CD ABE CE ABD DE ABC
15
Block

• We shall select candidates with almost the same
  eyesight level, and other gene conditions looking
  similar.
• We will make some preparation for the experiment
  to make candidates have similar mental
  conditions.

ONE BLOCK ONE PERSON
16
Methodology sum-up

• Two levels fractional factorial design
• 5 factorials 2(5-1)16 one block one person

Factors LOW LEVEL HIGH LEVEL
Color(latter) PURPLE YELLOW
Color(background) BLUE RED
Shape C E
Contrast 30 70
Luminance 50 90
Response 12 DEGREES 12 DEGREES
17
Preparation Performing

• DOE

18
Preparation

• Response variable
• Vision Capability
• Known Study
• Color
• Inverse colors increase the Vision Capability
• Colors with long wavelength decrease the VC

19
Preparation

• Known Study
• Shape
• Generally, Shape E is easier to identify than
  C.
• Contrast
• High contrast lead to high visual capacity.
• Luminance
• In a certain range, High luminance level
  increase the visual capacity.

20
Preparation

• Factor Level Code

NO. Factors Actual Level Coded Level
1 Color (Foreground) Yellow -1
1 Color (Foreground) Purple 1
2 Color (Background) Red -1
2 Color (Background) Blue 1
3 Shape C -1
3 Shape E 1
4 Contrast 30 -1
4 Contrast 70 1
5 Luminance 50 -1
5 Luminance 90 1
21
Preparation

• Held-constant factors
• Candidates Physical Conditions
• Distance from Vision Chart
• Light Color Luminance etc.

22
Experiment Performing

• Experiment Environment
• ZJ12510A
• Vision Chart
• Made by PowerPoint
• 8 Versions with different Foreground,
• Background color, and Symbol Shape
• Presented on the Computer Monitor
• in order to change the contrast and luminance
• Distance from Vision Chart 5 meters.

23
Data Analysis

• DOE

24
DATA ANALYSIS
Fractional factorial experiment
Var Factors Actual Level Coded Level
A Luminance 50 -1
A Luminance 90 1
B Color (Background) Red -1
B Color (Background) Blue 1
C Color (Foreground) Yellow -1
C Color (Foreground) Purple 1
D Shape C -1
D Shape E 1
E Contrast 30 -1
E Contrast 70 1
25
DATA ANALYSIS
Fractional factorial experiment
Normal probability plot
26
DATA ANALYSIS
Fractional factorial experiment
Term Effect Coef SE Coef T
P Constant 10.1719 0.05842 174.13
0.000 Block 1 0.0781 0.10118 0.77
0.444 Block 2 -0.1094 0.10118
-1.08 0.285 Block 3 -0.2969 0.10118
-2.93 0.005 A 0.2187 0.1094 0.05842
1.87 0.068 B -1.7813 -0.8906
0.05842 -15.25 0.000 C -0.9688 -0.4844
0.05842 -8.29 0.000 D 0.5938
0.2969 0.05842 5.08 0.000 E 0.0313
0.0156 0.05842 0.27 0.790 AB -0.0313
-0.0156 0.05842 -0.27 0.790 AC
-0.0938 -0.0469 0.05842 -0.80 0.427 AD
-0.0312 -0.0156 0.05842 -0.27 0.790 AE
0.0313 0.0156 0.05842 0.27 0.790 BC
-1.2188 -0.6094 0.05842 -10.43 0.000 BD
0.0937 0.0469 0.05842 0.80 0.427 BE
-0.2187 -0.1094 0.05842 -1.87 0.068 CD
0.1563 0.0781 0.05842 1.34
0.188 CE -0.1562 -0.0781 0.05842 -1.34
0.188 DE 0.0312 0.0156 0.05842
0.27 0.790 S 0.467336 R-Sq 91.15
R-Sq(adj) 87.62
ANOVA
27
DATA ANALYSIS
Fractional factorial experiment
Model check
28
DATA ANALYSIS
Fractional factorial experiment

• Further analysis
• Factor A and E are not significant
• Ignore A and E,
• Three factor two level and two replicates and
  four blocks full factorial design
• background color
• letter color
• letter shape
• four blocks four people

Full factorial experiment
29
DATA ANALYSIS
Full factorial experiment
30
DATA ANALYSIS
Full factorial experiment
Term Effect Coef SE Coef T
P Constant 10.1719 0.05936 171.35
0.000 Block 1 0.0781 0.10282 0.76
0.451 Block 2 -0.1094 0.10282
-1.06 0.292 Block 3 -0.2969 0.10282
-2.89 0.006 B -1.7813 -0.8906 0.05936
-15.00 0.000 C -0.9688 -0.4844
0.05936 -8.16 0.000 D 0.5938 0.2969
0.05936 5.00 0.000 BC -1.2188
-0.6094 0.05936 -10.27 0.000 BD 0.0937
0.0469 0.05936 0.79 0.433 CD
0.1563 0.0781 0.05936 1.32 0.194 BCD
0.0312 0.0156 0.05936 0.26 0.793 S
0.474901 R-Sq 89.24 R-Sq(adj) 87.21
31
DATA ANALYSIS
Full factorial experiment
32
DATA ANALYSIS
Full factorial experiment
33
DATA ANALYSIS
Full factorial experiment

• From the analysis, we get such conclusion
• For background color warm color (red) will lead
  to better performance.
• For letter color warm color (yellow) will lead
  to better performance.
• For color interaction inverse color will to
  better performance
• blue/ yellow and red/ purple is better than
  blue/purple and red/yellow
• For letter shape E shape will lead to better
  performance than C shape.
• For other factors under given condition not
  significant.

34
DATA ANALYSIS

• Check list to improve vision capability
• Red background color, purple letter color, E
  shape
• Inverse color pairs do better
• Warm color improve performance

35
TO BE IMPROVED

• Why other factors are not significant?
• luminance and contrast
• caused by device
• Response level is not very sensitive.
• Design method
• not linear, center points?
• not serious.

36
Reference

• ????????CRT???????
• ??????? ?????1995?1??1?
• VDT????????????????????
• ??? ??? ???? 2008 31(2)
• ???LogMAR????C???????????????
• ???, ???????2007 20(4)
• ?????????
• ???,???????? 2005.7
• ??????????????????
• ?? ??, ???????? 2007 28(7)

37
Acknowledgement

• Dr. Kaibo Wang
• TA Song XU
• People involved in the experiment
• Other team who help us in project

38
Thank You !