Abstract
The lightness of equiluminance regions of a light source appears different due to distinct properties of surrounding surfaces (lightness induction). For example, for a bar-grating stimulus in relatively low spatial-frequency, the same gray bars are perceived lighter when they are flanked by black bars, compared to those flanked by white bars (lightness contrast; Helson, 1963). However, the direction of this induction is reversed for a high spatial-frequency grating such that the gray among black bars appears darker (lightness assimilation), suggesting that the lightness is affected by stimulus spatial-frequency. More recently, Rudd (2010) also reported that the direction of lightness induction (contrast vs. assimilation) of a disk depended on the width and luminance of the ring surrounding the disk. We simulated these experiments with a biophysical retinal model and obtained qualitatively consistent results to the psychophysical data. For the stimulus conditions in which lightness contrast was observed, the model shows the classic center-surround lateral inhibition behavior that enhances signal contrast at the edge between neighboring surfaces. However, for a high spatial-frequency stimulus as in Helson (1963), the spatial high-frequency cutoff characteristics of the retinal MC pathway limits the effect of lateral inhibition such that the contrast enhancement occurs globally over the entire stimulus surface rather than at each edge between bars: thus, the signals of gray bars assimilate to those of the flanking bars. Also, a low and narrow ring as in Rudd (2010) fails to induce lateral inhibition and the signals at the edge of the disk assimilate to those of the ring. While lightness induction is not yet fully understood and in particular the mechanism for assimilation is yet of an unknown nature (Fiorentini 2004), our work would be the first to point out to retinal lateral inhibition as a starting point for the lightness contrast - assimilation phenomena.
Meeting abstract presented at VSS 2015