Probabilistic framework for person tracking on embedded distributed smart cameras

Tracking individuals is a prominent application in such domains like surveillance or smart environments. This paper provides a development of a multiple camera setup with disjointed view that tracks moving persons in a site. It focuses on a probabilistic modeling which utilizes the discriminative observed features such as person's appearance, and her/his possible pathways for the estimation of the unobserved identity. Each camera evaluates the difference between its locally extracted generative model with the corresponding received models from its neighbors to find the correspondence between prior and recent identified persons. The linear interpolation is applied to combine the observed features. In conjunction with this efficient probabilistic framework, a novel system on chip design for an FPGA-based smart camera is developed. It provides a hardware/software co-design architecture to achieve the real-time performance inside the smart camera. The functionality of the proposed system is evaluated through a realistic case study.