ROBUST FACE RECOGNITION USING AAM AND GABOR FEATURE VECTORS

摘要

A face recognition system based on an AAM and gabor feature vectors is provided to detect face feature points approximately by using the AAM, use the detected face feature point as an initial value, and detect face feature points in detail based on gabor jet similarity. A face region is detected in a model input image. The detected face region is normalized. By using an AAM(Active Appearance Model), an approximate face feature point is detected. By using gabor jet similarity, a detail face feature point is detected. Similarity with a registered gabor feature vector is measured. After that, it is determined whether to allow/reject.