Purchase this article with an account.
Mark Nawrot, Breanna Thompson, Shanda Lauer; Temporal properties of persistence and change in perceived depth from motion parallax. Journal of Vision 2018;18(10):125. doi: 10.1167/18.10.125.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Successful navigation by a moving observer requires rapid integration of visual depth information to construct a quickly updated 3D model of the environment. While unambiguous depth from motion parallax (MP) is generated with brief presentations (~30 msec), temporal properties for persistence and updating are unknown. Ten observers reported perceived depth (2AFC) upon viewing computer-generated random-dot MP stimuli. Stimulus presentation used a 120 Hz CRT with stimulus timing verified with an independent 20 MHz clock. MP stimuli made two oscillations, with each of the 4 lateral translations having the same duration (t). To maintain a consistent depth depiction, the direction of local stimulus dot movement reversed with each reversal in stimulus translation. Duration, t, was varied in two interleaved staircases, one for each initial direction of stimulus translation. Condition 1 confirmed that observers accurately recover unambiguous depth from MP with brief presentations (t = 34msec). Condition 2 found that observers require a longer duration (t = 93msec) to determine whether the last stimulus translation depicted MP depth, or was flat (no local dot movements). Condition 3 introduced a blank delay (duration = t) before the last translation in which observers made the same depth/no depth discrimination, which reduced the necessary presentation duration (t = 76msec). Condition 4 revealed that the discrimination of an MP depth reversal during the last stimulus translation requires a brief (t = 47msec) presentation. These results suggest that depth from MP can be recovered and updated in very brief temporal intervals, which is likely a useful property for an observer who is often shifting gaze during translation through a cluttered environment. However, these internal depth models appear to persist for much longer in the absence of depth information, which is perhaps useful in the maintenance of a consistent depth interpretation during blinks or brief obstructions of fixation.
Meeting abstract presented at VSS 2018
This PDF is available to Subscribers Only