Morphological algorithms for face detection

This paper presents an efficient method for video-based face detection using MRF (Markov random field) and morphology. First, MRF theory is used to extract the moving objects. Then, the top-hat filtering, bottom-hat filtering and watershed segmentation algorithms are employed to segment the image. Finally, the face region is characterized by elliptical shape as an ellipse can approximate the shape of a face. The experimental results show that this approach is very efficient and robust.