Estimation and interpretation of discontinuities in optical flow fields

Abstract

A systematic categorization of an adaptively determined local estimate for an OF-vector allows to detect non-local 'walls of discontinuities' which establish an essentially closed boundary around images of objects whose motion differs from that of foreground and background. The estimation process has been refined to the point where occasionally observable - initially counter-intuitive - discontinuity blobs inside a region corresponding to the image of a moving object found an acceptable explanation. Segmentation results obtained on this basis for different real-world image sequences will be used to illustrate the approach.