2003 年 22 巻 1 号 p. 108-114
Experimental results of perceptual grouping and apparent motion imply that perceptual information on similarities (or differences) in different perceptual dimensions are summed up. Linear regression analysis indicated that the obtained effects of multiple perceptual differences were simple weighted sums of effects of differences in individual perceptual dimensions. A perceptual state-space model is proposed. In the model, each dimension corresponds to each perceptual dimension: hue, brightness, shape, size or position, and a point represents a perceptual state. Perceptual grouping and apparent motion are assumed to occur between pairs of the closest two points in this state space. In this state space, city-block metric rather than Euclidean metric is effective. This means perceptual dimensions are separable with each other. Perceptual information is supposed to be synthesized after information-processing in individual perceptual dimensions.