Detecting Occlusions as an Inverse Problem

8 years 1 months ago

Download vision.ucla.edu

Occlusions generally become apparent when integrated over time because violations of the brightnessconstancy constraint of optical ﬂow accumulate in occluded areas. Based on this observation, we propose a variational model for occlusion detection that is formulated as an inverse problem. Our forward model adapts the brightness constraint of optical ﬂow to emphasize occlusions by exploiting their temporal behavior, while spatio-temporal regularizers on the occlusion set make our model robust to noise and modeling errors. In terms of minimization, we approximate the resulting variational problem by a sequence of convex optimizations and develop eﬃcient algorithms to solve them. Our experiments show the beneﬁts of the proposed formulation, both forward model and regularizers, in comparison to the state-of-the-art techniques that detect occlusion as the residual of optical-ﬂow estimation. Keywords occlusion detection, video processing

Virginia Estellers, Stefano Soatto

Real-time Traffic