摘要：This article investigates the problem of lightness perception. To clarify the role of depth inlightness perception two current models—the albedo hypothesis and the coplanar-ratiohypothesis—are discussed. To compare them the strength of the simultaneous-lightnesscontrast(SLC) illusion was investigated as a function of three-dimensional (3D) configurationsof the test and background squares. In accordance with both hypotheses thechanges in the depth arrangements of the test and background squares should result inchanges in the illusory effect. However, the reasons for and the directions of these changesshould be different. Five different types of 3D configurations were created in which thetest squares were tilted at different angles to the background squares. A virtual-realitytechnique was used to present stereo pairs of different 3D configurations. Thirty-sevenobservers took part in the experiment. The method of constant stimuli was used to obtainpsychometric functions. The displacements of these functions for 3D configurations incomparison with the 2D configuration allowed the estimation of illusion strength. Theanalysis of individual values of illusion strength revealed two groups of subjects. For thefirst group (38% of all participants) the strength changed insignificantly depending onthe 3D configurations. For the second group (62% of all participants) significant differenceswere obtained for those configurations in which the test and background squareswere perceived as differently illuminated. The changes in the SLC illusion strength for thesecond group were consistent with predictions made by the albedo hypothesis. Thus, itseems that the perceived illumination of a surface should be considered the main parameterfor lightness estimations in 3D scenes.
