3-D model-based tracking of humans in action: a multi-view approach

We present a vision system for the 3-D model-based t r acking of unconstrained human movement. Using image sequences acquired simultaneously f r om multiple views, we recover the 3-D body pose at each time instant without the use of markers. The pose-recovery problem is formulated a s a s e arch problem and entails nding the pose parameters of a graphical human model whose synthesized a p p earance i s m o s t similar to the actual appearance o f t h e r eal human in the multi-view images. The models used for this purpose are a c quired f r om the images. We use a decomposition approach and a best-rst technique to search through the high dimensional pose parameter space. A r obust variant of chamfer matching is used a s a fast similarity measure b etween synthesized and real edge images. We present initial tracking results from a large new Humans-InAction (HIA) database containing more than 2500 frames in each of four orthogonal views. They contain subjects involved in a variety of activities , of various degrees of complexity, ranging from the more simple one-person hand waving to the challenging two-person close interaction in the Argentine Tango.