Automatic Facial Feature Points Detection with SUSAN Operator
Haiyuan Wu · 2001
This paper proposes a method for detecting facial features and extracting its points from color images. A perceptually uniform chromatic system is used for representing color information in order to increasing the robustness of the extraction of the skin color region and the hair color regions. The face and the regions containing facial features are extracted from the image by applying the integral projection method, which makes use of both the color information (the skin color and the hair color) and the edge information (the strength and the orientation). The facial feature points are then detected from each region containing facial feature with the so called SUSAN corner detector. The proposed system is confirmed to be effective and robust to deal with the face images with various translations, slightly rotations and changes of facial expressions. Keyword: Facial feature point detection, Integral projection, SUSAN corner detector, Color information, Edge information