A new approach for motion characterization in image sequences is presented. It relies on the probabilistic modeling of temporal and scale co-occurrence distributions of local motion-related measurements directly computed over image sequences. Temporal multiscale Gibbs models allow us to handle both spatial and temporal aspects of image motion content within a unified statistical framework. Since this modeling mainly involves the scalar product between co-occurrence values and Gibbs potentials, we can formulate and address several fundamental issues: model estimation according to the ML criterion (hence, model training and learning) and motion classification. We have conducted motion recognition experiments over a large set of real image sequences comprising various motion types such as temporal texture samples, human motion examples, and rigid motion situations.
Motion recognition using nonparametric image motion models estimated from temporal and multiscale co-occurrence statistics
2003-12-01
Ieee Transactions On Pattern Analysis And Machine Intelligence (0162-8828) (Ieee Computer Soc), 2003-12 , Vol. 25 , N. 12 , P. 1619-1624
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
DDC: | 629 |
Non Parametric Motion Recognition Using Temporal Multiscale Gibbs Models
British Library Conference Proceedings | 2001
|Learned Temporal Models of Image Motion
British Library Conference Proceedings | 1998
|Qualitative Recognition of Motion Using Temporal Texture
British Library Online Contents | 1992
|Temporal co-occurrence matrix approaches to motion analysis [5298-26]
British Library Conference Proceedings | 2004
|