Sunghoon Kim, Hoon Lee, S. Park, K. Choi, and J. Hwang (Korea)
Face detection, video surveillance, moving object extraction, skin color filtering, SVM
In this paper, we present a framework to detect multiple faces with different sizes for surveillance applications, which is based on a fast moving object extraction approach followed by a skin color filtering and support vector machines. More specifically, moving objects are extracted first by using image difference between two consecutive frames and a simple object filling algorithm. Then, a skin color filtering combined with an eye detection algorithm is applied to locate possible face regions. Lastly, support vector machines are used to detect faces, which is applicable to detect faces with different sizes. Experimental results are provided to show that the proposed framework is applicable for surveillance applications.
Important Links:
Go Back
