6 June 2002 Image processing analysis of traditional Gestalt vision experiments
Author Affiliations +
Proceedings Volume 4421, 9th Congress of the International Colour Association; (2002) https://doi.org/10.1117/12.464559
Event: 9th Congress of the International Color Association, 2001, Rochester, NY, United States
In the late 19th century, the Gestalt Psychology rebelled against the popular new science of Psychophysics. The Gestalt revolution used many fascinating visual examples to illustrate that the whole is greater than the sum of all the parts. Color constancy was an important example. The physical interpretation of sensations and their quantification by JNDs and Weber fractions were met with innumerable examples in which two 'identical' physical stimuli did not look the same. The fact that large changes in the color of the illumination failed to change color appearance in real scenes demanded something more than quantifying the psychophysical response of a single pixel. The debates continues today with proponents of both physical, pixel-based colorimetry and perceptual, image- based cognitive interpretations. Modern instrumentation has made colorimetric pixel measurement universal. As well, new examples of unconscious inference continue to be reported in the literature. Image processing provides a new way of analyzing familiar Gestalt displays. Since the pioneering experiments by Fergus Campbell and Land, we know that human vision has independent spatial channels and independent color channels. Color matching data from color constancy experiments agrees with spatial comparison analysis. In this analysis, simple spatial processes can explain the different appearances of 'identical' stimuli by analyzing the multiresolution spatial properties of their surrounds. Benary's Cross, White's Effect, the Checkerboard Illusion and the Dungeon Illusion can all be understood by the analysis of their low-spatial-frequency components. Just as with color constancy, these Gestalt images are most simply described by the analysis of spatial components. Simple spatial mechanisms account for the appearance of 'identical' stimuli in complex scenes. It does not require complex, cognitive processes to calculate appearances in familiar Gestalt experiments.
© (2002) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
John J. McCann, John J. McCann, } "Image processing analysis of traditional Gestalt vision experiments", Proc. SPIE 4421, 9th Congress of the International Colour Association, (6 June 2002); doi: 10.1117/12.464559; https://doi.org/10.1117/12.464559

Back to Top