PERCEPTION OF SURFACE OF STRUCTURES OBEY THE PRINCIPLE OF GENERIC IMAGE SAMPLING IN MONKEYS.
T.Uka, H.Tanaka, M.Kato, and I.Fujita
Purpose. Humans perceive neon-color spreading in the Redies-Spillmann figure (R-S figure, a red cross with a white bar radiating from the end of each limb of the cross on a black background), and perceive two crossing bars in a cross with disparity added to the horizontal limbs. This has been taken as evidence that the human visual system uses the principle of generic image sampling as computational rule to determine 3-D surface structures from 2-D retinal images (Nakayama & Shimojo, 1992). We investigated whether this rule can be also applied to perception in monkeys.
Methods. Exp.1, A monkey was trained to discriminate a red cross from a transparent red disk superimposed on a red cross. After extensive training of the monkey, we examined whether the learning effect was transferred to discrimination between two new test figures, the R-S figure and the R-S figure with gray instead of white bars (no color spreading was perceived in experiments with human participants). Exp.2, The monkey was trained to discriminate between crosses with crossed disparities and those with uncrossed disparities on the horizontal limb. We then examined whether the learning effect was transferred to discrimination between crosses segmented by real contours as a horizontal bar in front of a vertical bar and as a vertical bar in front of a horizontal bar. The test figures were presented without disparity cues.
Results. The learning effect was transferred to discrimination between the test figures in both experiments.
Conclusions. The results indicate that the monkey perceives the neon-color spreading illusion and the depth structure of two bars as humans do; i.e. the principle of generic image sampling can be applied to monkeys.