Abstract
Many cues have been shown to be optimally combined for depth perception, including, in our own hands, realistically rendered shading and binocular disparity. Almost no studies, however, have explored the timing of cue combination. We therefore do not know whether particular cues are processed faster than others when both are present in a stimulus.
Here we use cue-conflict stimuli with forward and backward masking to explore how temporal constraints influence the combination of monocular shading and stereo disparity cues. Accurately rendered stimuli, containing shape from shading and binocular disparity were created, depicting a pair of convex surfaces (scaled circular cosine in depth), presented side by side. Observers were asked to indicate whether the left or right convex surface stood out less in depth. By introducing a cue-conflict between the shading and disparity cues in one of the convexities, we were able to track the relationship between the point of subjective equality (PSE) and the extent of conflict between cues. At the shortest presentations (25 msec), PSEs were based upon the coarse shading cue. As presentation times increased there was a shift towards the slower, but more reliable stereo cue (stimulus durations >100 msec).
In summary, we present the first evidence, from a cue-combination study, that supports the argument that shading cues alone can provide a quick-and-dirty estimate of depth and shape judgements. However, we also show that beyond around 100 msec, shading cues are vetoed in favour of the more reliable disparity cues.
EPSRC UK, (EP/G038708/1).