Joint Bayesian tracking of head location and pose from low-resolution video

This paper presents a visual particle filter for jointly tracking the position of a person and her head pose. The resulting information may be used to support automatic analysis of interactive people behavior, by supporting proxemics analysis and providing dynamic information on focus of attention. A pose-sensitive visual likelihood is proposed which models the appearance of the target on a key-view basis, and uses body part color histograms as descriptors. Quantitative evaluations of the method on the 'CLEAR'07 CHIL head pose' corpus are reported and discusssed. The integration of multi-view sensing, the joint estimation of location and orientation, the use of generative imaging models, and of simple visual matching measures, make the system robust to low image resolution and significant color distortion.