A novel framework for fast embedded face detection system

In this paper, we propose a novel design for one fast embedded face detection system, which can be applied in many real-time applications, such as teleconferencing, user interfaces, and security access control. Our framework includes 3 parts: one fast face detection method based on optimized AdaBoost algorithm with high speed and high detection rate, one SOC hardware framework to speed up detection operations and one software distribution strategy to optimize the memory sub-system. In embedded system domain, face detection is a great challenge because of its limited hardware source and low frequency. Our system prototype is built on the base of one SOC named ´Garfield´ to test our design. Experimental results illuminate the validity of our face detection system.