Non parametric motion recognition using temporal multiscale Gibbs models

Publisher Website
Google Scholar
Abstract
We present an original approach for non parametric mo- tion analysis in image sequences. It relies on the statistical modeling of distributions of local motion-related measure- ments computed over image sequences. Contrary to pre- viously proposed methods, the use of temporal multiscale Gibbs models allows us to handle in a unified statistical framework both spatial and temporal aspects of motion con- tent. The important feature of our probabilistic scheme is to make the exact computation of conditional likelihood func- tions feasible and simple. It enables us to straightforwardly achieve model estimation according to the ML criterion and to benefit from a statistical point of view for classification issues. We have conducted motion recognition experiments over a large set of real image sequences comprising vari- ous motion types such as temporal texture samples, human motion examples and rigid motion situations.

This publication has 14 references indexed in Scilit: