39 |
- mail to jyu-ui about the 1D color example |
- mail to jyu-ui about the 1D color example |
40 |
|
|
41 |
|
|
|
2003-08-26: |
|
|
- meeting with psa at 13:00 |
|
|
|
|
42 |
============= |
============= |
43 |
Past: |
Past: |
44 |
============= |
============= |
45 |
|
|
46 |
|
2003-08-26: |
47 |
|
- fillets meeting with psa and humppake |
48 |
|
- some ideas to think about: |
49 |
|
- show the nodes (without connections) before |
50 |
|
showing the full graph in each trial |
51 |
|
--> search time eliminated |
52 |
|
- color all nodes differently and interpolate |
53 |
|
the color over the connections --> |
54 |
|
should be even less ambiguous |
55 |
|
(probably not usable for showing the |
56 |
|
superiority of fillets) |
57 |
|
- conclusion: |
58 |
|
- experiments first |
59 |
|
|
60 |
|
- read the book "Robert M. Boynton: Human Color Vision, |
61 |
|
New York: Holt, Rinehart and Winston, 1979." Notes: |
62 |
|
- sensitivity regulation for lightness is well known: |
63 |
|
the size of pupil, pigment bleaching, and receptor |
64 |
|
sensitivity adapt the sensitivity |
65 |
|
- however, chroma sensitivity as a function of chroma range |
66 |
|
is not directly considered in the book |
67 |
|
- contours play a crucial role in color vision: |
68 |
|
colors are averaged over regions separated by contours |
69 |
|
- B cones do not participate in contour perception |
70 |
|
(there are very few B cones in the fovea) |
71 |
|
- colors whose R-G and R+G responses are identical |
72 |
|
are discriminated better if a small gap is inserted |
73 |
|
between the fields so as to enable contour perception |
74 |
|
|
75 |
|
|
76 |
2003-08-25: |
2003-08-25: |
77 |
- tried "is there one or two connected componens?" task |
- tried "is there one or two connected componens?" task |
78 |
- too hard for >12 nodes --> cannot measure scaling |
- too hard for >12 nodes --> cannot measure scaling |