Title: Color appearance of natural objects
1Color appearance of natural objects
- Thorsten Hansen, Sebastian Walter
- and Karl R. Gegenfurtner
Department of Psychology University of
Giessen Germany
2Stages of color vision
- cones 2. cone-opponent 3. higher-order
4. color -
mechanisms categories
retina retina/LGN
visual cortex
black white red green yellow blue orange pink gray
brown
S
M
L
3Stadien der Farbverarbeitung
4Motivation
Previous studies have investigated the color
appearance of homogeneous disks in the
isoluminant plane of DKL color space (e.g., De
Valois et al. 1997, Webster Miyahara 2000, De
Valois et al. 2000). These studies investigated
hue scaling for the unique hues RED, GREEN,
YELLOW, and BLUE in DKL color space. These
studies have shown that the cardinal axes do not
coincide with the unique hues. In particular, the
S-(LM) axis varies from violet to
yellow-green. Stimuli were always homogenous
patches.
5Motivation
In contrast to standard disk stimuli, natural
objects are characterized by a distribution of
different chromatic hues.
6Motivation
- Here we investigated two question
- Does the presentation of colors at different
luminance values alters the location of the
unique hue? - Does the use of natural objects with a
distinctive object color (e.g., a banana)
influence the color appearance ?
7DKL color space
Derrington Krauskopf Lennie
Cone-opponent color space two chromatic axes
(L-M and S-(LM)), one luminance axis (LM).
8DKL cone-opponent axes vs. Hering's opponent
colors
Ewald H. Hering 18341918
9Methods
Stimulus creation translate color distribution
of the objects in the isoluminant plane of DKL
space.
10banana space
11Methods
Observers rated the appearance of the unique hues
(RGYB) for stimuli at 36 equally spaced chromatic
direction in the isoluminant plane of DKL color
space.
12Cone contrast in DKL space
13Methods
Observers rated the appearance of the unique hues
(RGYB) on a scale from 0 to 8.
R B G Y
Stimuli at each color directions were shown twice
in randomized order in each block. Five blocks
were run for each observer, resulting in a total
of 10 hue scaling for each chromatic direction.
14Methods
- Natural fruit stimuli (banana, salad, carrot) are
compared against - colored disks of homogeneous color
- outline shapes of homogeneous color
- at 5 different luminance values.
- N3 naïve observers participated in the study,
with normal color vision as assessed by Ishihara
pseudoisochromatic plates.
15Results
Color appearance of a standard disk stimulus.
16Results Variation of brightness.
17Results Natural objects
18Results Single subject data
19Results Comparison with outline shapes
20Results Other fruit salad.
21Summary
- In agreement with earlier studies we found that
- The cardinal axes do not coincide with the unique
hues - This is true for stimuli at a large range of
different brightness values.
?
22Summary
The effect is smaller for outline shapes compared
to real images with a variety of hues.
The present study complements other work of our
group using natural objects where we have
investigated color discrimination (VSS 2005, ECVP
2005) and color memory (ECVP 2004).
23Conclusion
The distribution of hues in natural objects
influences color appearance and needs to be taken
into account for a full characterization of color
appearance.