Limits...
Contribution of a luminance-dependent S-cone mechanism to non-assimilative color spreading in the watercolor configuration.

Kimura E, Kuroki M - Front Hum Neurosci (2014)

Bottom Line: When the luminance condition was reversed and the IC contrast was greater than the OC contrast (lower IC luminance condition), the color spreading was non-assimilative and yellowish.When the color spreading was analyzed in terms of cone-opponent excitations, the results were consistent with the interpretation that the color spreading is explainable by a combination of chromatic diffusion from the IC and chromatically opponent induction from the OC.These findings provided several constraints on possible visual mechanisms underlying the watercolor effect.

View Article: PubMed Central - PubMed

Affiliation: Department of Psychology, Faculty of Letters, Chiba University Chiba-shi, Japan.

ABSTRACT
In the watercolor configuration composed of wavy double contours, both assimilative and non-assimilative color spreading have been demonstrated depending on the luminance conditions of the inner and outer contours (IC and OC, respectively). This study investigated how the induced color in the watercolor configuration was modulated by combinations of the IC and the OC color, particularly addressing non-assimilative color spreading. In two experiments, the IC color was fixed to a certain color and combined with various colors selected from a hue circle centered at the background white color. Color spreading was quantified with a chromatic cancelation technique. Results showed that both the magnitude and the apparent hue of the color spreading were largely changed with the luminance condition. When the IC contrast (Weber contrast of the IC to the background luminance) was smaller in size than the OC contrast (higher IC luminance condition), the color spreading was assimilative. When the luminance condition was reversed and the IC contrast was greater than the OC contrast (lower IC luminance condition), the color spreading was non-assimilative and yellowish. When the color spreading was analyzed in terms of cone-opponent excitations, the results were consistent with the interpretation that the color spreading is explainable by a combination of chromatic diffusion from the IC and chromatically opponent induction from the OC. The color spreading in the higher IC luminance condition mainly reflected the chromatic diffusion by both (L-M) and S cone-opponent mechanisms. The non-assimilative color spreading in the lower IC luminance condition mostly reflected S-cone mediated opponent induction and the contribution of -S inducing mechanisms was differentially large. These findings provided several constraints on possible visual mechanisms underlying the watercolor effect.

No MeSH data available.


Color direction and shift size of the cancelation settings as a function of the azimuth of the OC color. (A,C) show results obtained in the higher IC luminance condition; (B,D) show results obtained in the lower IC luminance condition. Upper panels show the results of color direction analysis, whereas lower panels show the results of shift size analysis. The color direction was defined as the angle from the +u′ axis to the direction of the mean setting. The direction is opposite to that of the perceived color. The shift size was defined as the vector length (i.e., Euclidean distance) from the background white to the mean setting, expressed in the percentage to the Euclidean distance of the inducing contour colors (0.05). Circle, triangle, and square symbols respectively show results in the red IC, orange IC, and achromatic IC conditions. Error bars show ±1 SEM across observers. In the figures of color direction analysis (A,B), important directions were also shown by horizontal dashed lines [opposite directions of the orange and red colors, +S/(L+M), and +L/(L+M) axis]. In the figures of shift size analysis (C,D), the opposite directions of the orange and red colors, and the +S/(L+M) axis were shown (from left to right) as colored vertical dashed lines.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4260510&req=5

Figure 5: Color direction and shift size of the cancelation settings as a function of the azimuth of the OC color. (A,C) show results obtained in the higher IC luminance condition; (B,D) show results obtained in the lower IC luminance condition. Upper panels show the results of color direction analysis, whereas lower panels show the results of shift size analysis. The color direction was defined as the angle from the +u′ axis to the direction of the mean setting. The direction is opposite to that of the perceived color. The shift size was defined as the vector length (i.e., Euclidean distance) from the background white to the mean setting, expressed in the percentage to the Euclidean distance of the inducing contour colors (0.05). Circle, triangle, and square symbols respectively show results in the red IC, orange IC, and achromatic IC conditions. Error bars show ±1 SEM across observers. In the figures of color direction analysis (A,B), important directions were also shown by horizontal dashed lines [opposite directions of the orange and red colors, +S/(L+M), and +L/(L+M) axis]. In the figures of shift size analysis (C,D), the opposite directions of the orange and red colors, and the +S/(L+M) axis were shown (from left to right) as colored vertical dashed lines.

Mentions: To quantify the changes in the induced color further, the color direction and shift size of each observer's cancelation settings were calculated for a u′v′ chromaticity diagram and were then averaged across observers (Figure 5) (See Kimura and Kuroki, 2014 for details of the analysis). The color direction was defined as the angle from the +u′ axis to the direction of the mean cancelation setting. The color direction would be 194.0° in the red IC condition and 226.3° in the orange IC condition if the induced color had the same hue as the IC color. The magnitude of color spreading was quantified with the shift size defined as the distance to the mean cancelation setting from the background white point, expressed in the percentage to the Euclidean distance of the inducing contour colors (0.05). The result of the color direction analysis in the higher IC luminance condition (Figure 5A) showed that the color directions of the cancelation settings were similar to those of the IC color in the red IC and the orange IC condition (circle and triangle symbols, respectively), which denotes assimilative color spreading. A One-Way repeated-measures ANOVA showed a significant main effect of the OC color in the red IC and the orange IC condition [F(7, 28) = 4.381, p = 0.0022 and F(7, 28) = 2.611, p = 0.0330, respectively]. Multiple comparison tests (Ryan, α = 0.05) showed that the angle of the color direction was larger when the azimuth of the OC color was 284.0° than when it was 59.0, 148.9, or 194.0° in the red IC condition. However, no significant difference was found from multiple comparison tests conducted in the orange IC condition. Consequently, the deviation in the color direction from that of the IC color when the azimuth of the OC color was around 270° (Figure 4A) was statistically significant in the red IC condition.


Contribution of a luminance-dependent S-cone mechanism to non-assimilative color spreading in the watercolor configuration.

Kimura E, Kuroki M - Front Hum Neurosci (2014)

Color direction and shift size of the cancelation settings as a function of the azimuth of the OC color. (A,C) show results obtained in the higher IC luminance condition; (B,D) show results obtained in the lower IC luminance condition. Upper panels show the results of color direction analysis, whereas lower panels show the results of shift size analysis. The color direction was defined as the angle from the +u′ axis to the direction of the mean setting. The direction is opposite to that of the perceived color. The shift size was defined as the vector length (i.e., Euclidean distance) from the background white to the mean setting, expressed in the percentage to the Euclidean distance of the inducing contour colors (0.05). Circle, triangle, and square symbols respectively show results in the red IC, orange IC, and achromatic IC conditions. Error bars show ±1 SEM across observers. In the figures of color direction analysis (A,B), important directions were also shown by horizontal dashed lines [opposite directions of the orange and red colors, +S/(L+M), and +L/(L+M) axis]. In the figures of shift size analysis (C,D), the opposite directions of the orange and red colors, and the +S/(L+M) axis were shown (from left to right) as colored vertical dashed lines.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4260510&req=5

Figure 5: Color direction and shift size of the cancelation settings as a function of the azimuth of the OC color. (A,C) show results obtained in the higher IC luminance condition; (B,D) show results obtained in the lower IC luminance condition. Upper panels show the results of color direction analysis, whereas lower panels show the results of shift size analysis. The color direction was defined as the angle from the +u′ axis to the direction of the mean setting. The direction is opposite to that of the perceived color. The shift size was defined as the vector length (i.e., Euclidean distance) from the background white to the mean setting, expressed in the percentage to the Euclidean distance of the inducing contour colors (0.05). Circle, triangle, and square symbols respectively show results in the red IC, orange IC, and achromatic IC conditions. Error bars show ±1 SEM across observers. In the figures of color direction analysis (A,B), important directions were also shown by horizontal dashed lines [opposite directions of the orange and red colors, +S/(L+M), and +L/(L+M) axis]. In the figures of shift size analysis (C,D), the opposite directions of the orange and red colors, and the +S/(L+M) axis were shown (from left to right) as colored vertical dashed lines.
Mentions: To quantify the changes in the induced color further, the color direction and shift size of each observer's cancelation settings were calculated for a u′v′ chromaticity diagram and were then averaged across observers (Figure 5) (See Kimura and Kuroki, 2014 for details of the analysis). The color direction was defined as the angle from the +u′ axis to the direction of the mean cancelation setting. The color direction would be 194.0° in the red IC condition and 226.3° in the orange IC condition if the induced color had the same hue as the IC color. The magnitude of color spreading was quantified with the shift size defined as the distance to the mean cancelation setting from the background white point, expressed in the percentage to the Euclidean distance of the inducing contour colors (0.05). The result of the color direction analysis in the higher IC luminance condition (Figure 5A) showed that the color directions of the cancelation settings were similar to those of the IC color in the red IC and the orange IC condition (circle and triangle symbols, respectively), which denotes assimilative color spreading. A One-Way repeated-measures ANOVA showed a significant main effect of the OC color in the red IC and the orange IC condition [F(7, 28) = 4.381, p = 0.0022 and F(7, 28) = 2.611, p = 0.0330, respectively]. Multiple comparison tests (Ryan, α = 0.05) showed that the angle of the color direction was larger when the azimuth of the OC color was 284.0° than when it was 59.0, 148.9, or 194.0° in the red IC condition. However, no significant difference was found from multiple comparison tests conducted in the orange IC condition. Consequently, the deviation in the color direction from that of the IC color when the azimuth of the OC color was around 270° (Figure 4A) was statistically significant in the red IC condition.

Bottom Line: When the luminance condition was reversed and the IC contrast was greater than the OC contrast (lower IC luminance condition), the color spreading was non-assimilative and yellowish.When the color spreading was analyzed in terms of cone-opponent excitations, the results were consistent with the interpretation that the color spreading is explainable by a combination of chromatic diffusion from the IC and chromatically opponent induction from the OC.These findings provided several constraints on possible visual mechanisms underlying the watercolor effect.

View Article: PubMed Central - PubMed

Affiliation: Department of Psychology, Faculty of Letters, Chiba University Chiba-shi, Japan.

ABSTRACT
In the watercolor configuration composed of wavy double contours, both assimilative and non-assimilative color spreading have been demonstrated depending on the luminance conditions of the inner and outer contours (IC and OC, respectively). This study investigated how the induced color in the watercolor configuration was modulated by combinations of the IC and the OC color, particularly addressing non-assimilative color spreading. In two experiments, the IC color was fixed to a certain color and combined with various colors selected from a hue circle centered at the background white color. Color spreading was quantified with a chromatic cancelation technique. Results showed that both the magnitude and the apparent hue of the color spreading were largely changed with the luminance condition. When the IC contrast (Weber contrast of the IC to the background luminance) was smaller in size than the OC contrast (higher IC luminance condition), the color spreading was assimilative. When the luminance condition was reversed and the IC contrast was greater than the OC contrast (lower IC luminance condition), the color spreading was non-assimilative and yellowish. When the color spreading was analyzed in terms of cone-opponent excitations, the results were consistent with the interpretation that the color spreading is explainable by a combination of chromatic diffusion from the IC and chromatically opponent induction from the OC. The color spreading in the higher IC luminance condition mainly reflected the chromatic diffusion by both (L-M) and S cone-opponent mechanisms. The non-assimilative color spreading in the lower IC luminance condition mostly reflected S-cone mediated opponent induction and the contribution of -S inducing mechanisms was differentially large. These findings provided several constraints on possible visual mechanisms underlying the watercolor effect.

No MeSH data available.