This video demonstrates a real time implementation of FPGA Based Human Crowd Motion Classification System. It classifies the human crowd behavior using motion vectors statistical features and categorizes them into 'Panic/Fighting, Normal/Swift Walking, Running, Fast Running' etc. It also detects whether people are diverging or their motion momentum has changed in the scene. The hardware used is Spartan-6 ATLYS FPGA board and the over all application built uses Xilinx ISE, Vivado HLS, and Embedded Development Kit.
