Abstract
Drawings, paintings, photos, videos, and movies are acceptable, indeed sought after, 2-D stand-ins for 3-D scenes and make up a large and growing fraction of the total visual input of many modern humans. Yet they severely distort the visual signals available in the 3-D version. Binocular disparities, monocular occlusions, blur, motion parallax, and perspective are absent or altered. Still, viewers of these images are seldom bothered, or even notice, these distortions, implying an adaptation to the cues available in 2-D visual media. In order to assess this adaptation through its aftereffects, we used commercial movies in 2-D, 3-D, and reversed-stereo versions as adapting stimuli. Of particular interest were effects due to the absence of disparity cues and the inconsistency between perspective and disparity. We measured observers' sensitivity to depth using three tasks: (1) discrimination and (2) adjustment of the relative stereo-depth of two vertical lines in the presence of irrelevant depth-from-perspective cues, and (3) depth discrimination of random-dot stereograms. In discrimination tasks, observers reported the line or random-dot interval having the nearer depth value. In the adjustment procedure, observers set a perceptual depth match between the lines. Stereoacuity was higher after watching 2-D movies for all but one observer (total: 5) and all tasks. Perspective had an irregular effect across observers. Regular and reversed-depth 3-D movies did not affect stereoacuity or depth from perspective. Reading text under movie-viewing conditions improved stereoacuity, the same as movie-watching. We conclude that depth cue conflict is not essential for the improvement in stereo performance. Adaptation to near-zero disparities appears sufficient to raise sensitivity to subsequent non-zero disparities, with recovery to baseline extending over a period of approximately 30 minutes post-adaptation.
Meeting abstract presented at VSS 2017