Purchase this article with an account.
Kyungim Baek, Paul Sajda; A probabilistic network model of the influence of local figure-ground representations on the perception of motion. Journal of Vision 2005;5(8):842. doi: https://doi.org/10.1167/5.8.842.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Psychophysical experiments have shown that integration of motion signals, distributed across space, must be integrated with form cues, such as those associated figure-ground segregation. These experiments have led several to conclude that mechanisms exist which enable form cues to ‘veto’ or completely suppress ambiguous motion signals. We present a probabilistic network model in which local figure-ground representations encoded by direction-of-figure (Sajda and Finkel, 1995) modulate the degree of certainty of local motion signals. In particular, we consider the modulation at junctions where line terminators are defined as either intrinsic or extrinsic (Shimojo, Silverman, and Nakayama, 1989). The strength of local motion suppression at extrinsic terminators is a function of the belief in the local direction-of-figure, which is defined as the strength of the evidence for surface occlusion. Unlike previous studies/models investigating the influence of motion signals at terminators and occlusion cues (Grossberg, Mingolla, and Viswanathan, 2001; Lidén and Pack, 1999), our model directly exploits the uncertainties in the observations (i.e. figure-ground cues) leading to uncertainty in the inferred direction-of-figure, which for the case of terminators provides a smooth transition between intrinsic and extrinsic classes. Simulation results show that our model can account for the continuum of perceptual bias seen for motion coherence and perceived direction of motion in psychophysical experiments (McDermott, Weiss, and Adelson, 2001; Lidén and Mingolla, 1998).
This PDF is available to Subscribers Only