Fully automatic face detection and facial feature points extraction using local Gabor filter bank and PCA

Detecting facial feature points in images is a crucial stage for facial action interpretation tasks. This paper proposes a facial feature points extraction system based on local Gabor filter bank and principal component analysis (PCA). Usually, a Gabor filter bank is formed by 5 frequencies and 8 orientations. In this case, a lot of time is consumed while many information is useless. In this paper, we only apply 3 frequencies and 4 orientations to form a local Gabor filter bank and employ the PCA method to reduce the dimension further. Experimental results show that the local Gabor filter bank formed by 12 Gabor filters also performs very well in feature points extraction so that the efficiency of the system can be improved.