Abstract
Depth from binocular disparity is an important cue for segregating an object from its background, and potentially breaking camouflage. In this study we explored whether depth from disparity is hindered by luminance ‘camouflage’, where luminance edges on an object are not coincident with disparity-defined edges. We presented participants with random dot stereograms containing a binocular disparity-defined circle. In some stimuli, luminance edges were located at a smaller, or larger radius than that of the disparity-defined circle. We used a 2afc task, asking observers to decide which of two stimuli contained the greatest peak depth, and obtained thresholds and biases for depth perception. We found biases in the peak perceived depth that depended on the radius of the luminance-defined circle. When the radius of the luminance defined circle was smaller than the disparity-defined object, there was little effect on the perceived peak depth. However, if the luminance patch had a larger radius than the disparity defined circle, we found a significant perceptual bias, with the peak depth being underestimated and the threshold increased. We will present a simple model suggesting that peak depth is being estimated using the some form of average over the disparities contained within the luminance defined circle. Our results demonstrate that depth perception of an object is affected by simple manipulations of luminance edges. The interaction of luminance and disparity suggests a possible route to beat the camouflage breaking properties of disparity-defined depth.
Meeting abstract presented at VSS 2015