Purchase this article with an account.
Qasim Zaidi, Rocco Robilotto; Material identification for patterned 3-D objects. Journal of Vision 2004;4(8):344. doi: https://doi.org/10.1167/4.8.344.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Even though most objects in the world are neither uniform in color nor flat, lightness constancy has been studied for flat objects of uniform reflectance, while contrast constancy has been studied for variations in frequency, orientation and luminance of bars and gratings. We used real illuminants and 3-D objects under naturalistic viewing conditions to measure observers' abilities to identify patterned materials across illuminants. The stimuli were randomly crumpled papers printed with achromatic patterns with precisely controlled mean-reflectance and reflectance-contrast. Pairs of objects were placed inside two side-by-side compartments. Three of the objects (standards) were crumpled from the same material, while the material of the fourth (test) differed in either mean-reflectance or reflectance-contrast. The light on one compartment was four times the radiance on the other, both were lined with 40 gray levels. Two levels of mean-reflectance and two levels of reflectance-contrast were used for standard objects. On each trial, the observer was asked to identify the object with the different material. For each standard, the frequency of picking either object in the compartment that contained the test object, gives the psychometric function for reflectance discrimination within illuminants. The frequency of picking the correct test object gives the psychometric function for reflectance identification across illuminants. By comparing identification to discrimination thresholds, we tested whether reflectance identification is better for patterned objects because perceived contrasts are affected less than perceived brightness by the illuminant radiance. In two additional experiments, using the identical conditions, observers were asked to pick the material most dissimilar in contrast or brightness respectively. The reflectance identification data was fit with a model combining judgments of contrast and brightness dissimilarity with brightness adaptation.
This PDF is available to Subscribers Only