In support of crime prevention the human monitoring systems are widely deployed in companies, schools and elsewhere. There are number of effective methods to predict the migration route of a person in a crowd by using high-order particle filter and online-learning. There are method focused on group structure to improve tracking accuracy in a situation when the detection ranges of cameras overlap.
Papers:
M. Shiozuka, T. Yotsumoto, K. Takahashi, M. Nishiyama, T. Kawamura, and K. Sugahara, “Countermeasure to Human Recognition Error for Agent-based Human Tracking System,” 12th International Conference on Mobile Ubiquitos Computing, Systems, and Technologies (UBICOMM2018), pp. 65- 70.
M. Shiozuka, T. Yotsumoto, K. Takahashi, T. Kawamura, and K. Sugahara, “Implementation Example with Ultra-Small PCs for Human Tracking System Based on Mobile Agent Technologies,” 11th International Conference on Mobile Ubiquitos Computing, Systems, and Technologies (UBICOMM2017), pp. 73-78.
Y.J. Cho, S.A. Kim, J.H. Park, K. Lee, and K.J. Yoon, “Joint Person Re-identification and Camera Network Topology Inference in Multiple Camera,” arXiv:1710.00983, 2017.