A probabilistic reasoning approach to closed-room people monitoring

We propose a novel probabilistic reasoning approach to recognizing people entering and leaving a closed room in the work place or living environment. In our method people are represented using a low-level color feature, based on which optimal recognition is carried out by exploiting the temporal correlation among the sequence of observations of people. Optimality is achieved by maximizing a joint posterior probability. Experimental results are presented to show the efficacy of the proposed approach.