How Do Reliability and Timing Influence Cue-Combinations for Shading and Stereo-Disparity?

Abstract

Cue-conflict studies have been used with considerable success to examine the mechanisms underlying the integration of different sensory cues. Our recent studies have used these methods to examine the combination of stereo-disparity and shading cues in realistically rendered stimuli. The stimuli featured circular convexities with cyclopean shading and a stereo-cue based upon a densely dotted surface texture—other cues (shape outline and surface texture gradient) were controlled for. In our first experiment observers made depth discriminations while disparity noise was manipulated; we were able to demonstrate that for most observers cues are combined optimally—ie, PSEs and response variances shifted towards the more reliable channel. In a follow-on study we have examined how these cue-combinations are affected by presentation timing—with forwards and backwards masking. For most observers the adage that shading provides a quick and dirty mechanism for making depth judgements seems to be supported. Observers can make reliable depth judgements based upon shading when the difference in depth is relatively fine. However, for fine discriminations more time is required, and by this time the more accurate information from the stereo-channel has become available and is used.