Low Level Recognition of Human Motion Or How to Get Your Man Without Finding his Body Parts

The recognition of human movements such as walk ing running or climbing has been approached previ ously by tracking a number of feature points and either classifying the trajectories directly or matching them with a high level model of the movement A major di culty with these methods is acquiring and tracking the requisite feature points which are generally speci c joints such as knees or ankles This requires previous recognition and or part segmentation of the actor In this paper we show that the recognition of walking or any repetitive motion activity can be accomplished on the basis of bottom up processing which does not re quire the prior identi cation of speci c parts or clas si cation of the actor In particular we demonstrate that repetitive motion is such a strong cue that the moving actor can be segmented normalized spatially and temporally and recognized by matching against a spatiotemporal template of motion features We have implemented a real time system that can recognize and classify repetitive motion activities in normal gray scale image sequences