Detecting Humans under Partial Occlusion using Markov Logic Networks

Identifying humans under partial occlusion is a challenging problem in unconstrained scene understanding. In contrast to many existing works that model human appearance in isolation, we address this problem by studying the seman-tic context between human face and other body parts using Markov logic networks. By learning a set of probabilistic first-order logic rules that capture interactions between body parts under varying degrees of occlusion, and the relation-ship they share with the neighboring spatial windows, we obtain a graphical model representation of these instances to facilitate inference. We illustrate the efficacy of our method through experiments on standard human detection datasets, and an internally collected dataset with several occluding humans