Real-time action recognition in embedded vision systems

点击翻译按钮获取中文摘要

The proposed project consists then in the development of a video analysis system that can detect, in real-time, the occurrence of specific on-going people actions. The algorithms to be developed will subsequently be embedded into the Synopsys embedded vision platform. The objective is to have this intelligent module integrated into the next generation of smart video surveillance cameras. The prototypal system will be developed and tested at the University of Ottawa and deployed on the field during the final phase of the project. Performance of the system will be evaluated on the basis of the number of correctly detected occurring actions, the number of missed actions and the number of false alarms generated by the system. This action recognition system will be designed such that it can easily be mapped to an embedded architecture.