A Model for the Estimate of Local Velocity

Abstract

Motion sensitive cells in the primary visual cortex are not selective to velocity, but rather are directionally selective and tuned to spatiotemporal frequencies. This paper describes physiologically plausible theories for computing velocity from the outputs of spatiotemporally oriented filters and proves several theorems showing how to combine the outputs of a class of frequency tuned filters to detect local image velocity. Furthermore, it can be shown (Grzywacz and Yuille 1990) that the filters' combination may simulate “Pattern” cells in the middle temporal area (MT), while each filter simulates primary visual cortex cells. This suggests that MT's role is not to solve the aperture problem, but to estimate velocities from primary cortex information. The spatial integration that accounts for motion coherence may be postponed to a later cortical stage.