Hidden cues in random-line stereograms

Vernier acuity shows that the human visual system can interpolate the relative position of a feature with astonishing precision, of the order of a fraction of the distance between neighbouring photoreceptors in the fovea. The remarkable properties of this interpolation process have been the subject of many recent studies1–9. A major question that is still unanswered is whether the process is parallel—operating automatically at all times over a large part of the visual field—or serial—operating selectively at isolated locations only when judgments in the hyperacuity range are required (as in a forced choice task). Successful fusion of random-line stereograms having breaks in the vernier acuity range has been previously interpreted10 to suggest that the interpolation process underlying hyperacuity is parallel and preliminary to stereomatching. Here we demonstrate with computer experiments that vernier cues are not needed to solve the stereomatching problem posed by these stereograms, and we provide psychophysical evidence that human stereopsis probably does not use vernier cues alone to achieve fusion of these random-line stereograms.