Iroshini Gunasekera^1,2 (), Romina Abadi^1,3, Faruq Afolabi^1,3, Xue Teng^1,3, Robert S. Allison^1,3, Laurie M. Wilcox^1,2; ¹Centre for Vision Research, York University, ²Department of Psychology, York University, ³Department of Electrical Engineering and Computer Science, York University

Studies of size perception have tended to focus on the effect of distance and used simple 2D stimuli. Factors affecting the perceived size of volumetric 3D shapes have received less attention. Here, we evaluate the contribution of monocular and binocular depth information to size perception, during passive viewing and active interaction with the stimuli. Using virtual reality (VR) we presented a virtual shape-posting toy. Targets were 3D shapes (triangle, pentagon, square, and quatrefoil) and on each trial, one target appeared alongside a box with identically shaped slots. Participants indicated if the target was larger or smaller than the corresponding slot in a 2AFC task. Shape size varied according to the method of constants; the psychometric data were fit with cumulative normal distributions to compute JNDs and PSEs. We assessed the effects of stereopsis and object motion in two experiments. In Experiment 1 (N=28) the shapes were viewed monocularly or binocularly and were stationary or could be picked up and rotated. In Experiment 2 (N=26) we evaluated the effect of two types of motion (passive rotation vs active movement) and monocular vs. binocular viewing. In Experiment 2, positional data for both head movements and object trajectories were collected. We found that observers discriminated size accurately across all test conditions; PSEs showed no consistent bias. Overall, binocular judgements were more precise (smaller JNDs) than monocular judgements. Surprisingly, in the monocular conditions, discrimination performance worsened when participants interacted with the object. Analysis of target positions during trials showed that observers did not adopt different strategies depending on the depth cues available (e.g. they did not bring monocularly viewed objects closer). Our results underscore the importance of binocular depth information in perception of 3D object size, even when motion or active interaction could theoretically enhance size judgements.

Acknowledgements: This research was supported by grants RGPIN-2019-06694 to LW and ALLRP 570802–21 to RA and LW from the Natural Science and Engineering Council (Canada), and CF-REF VISTA funding to I. Gunasekera.