Learning to recognize visual dynamic events from examples

This paper describes a trainable and flexible system able to recognize visu al dynamic events, e.g. movements performed by different people, from a str eam of images taken by a fixed camera. Each event is represented by a featu re vector built from the spatio-temporal changes detected in the observed i mage sequence. The system neither attempts to recover the 3D structure nor assumes a prior model of the observed dynamic events. During training a sup ervisor identifies and labels the events of interest among those automatica lly detected by the system. At run time, previously unseen events are detec ted and classified on the basis of the available examples. Several experime nts on real images are reported and the benefits of using Support Vector Ma chines for performing effective classification from a relatively small numb er of labeled examples and for building noise tolerant representations are discussed. Preliminary results indicate that the proposed system can also b e applied with equally good results to the case in which the dynamic events are gestures performed by different people.