Face Recognition from 2D and 3D Images

Abstract

This paper proposes a feature-based face recognition system based on both 3D range data as well as 2D gray-level facial images. Feature points are designed to be robust against changes of facial expressions and viewpoints and are described by Gabor Wavelet filter in the 2D domain and Point Signature in the 3D domain. Localizing feature points in a new image is based on 3D-2D correspondence, their relative position and corresponding Bunch (covering a wide range of possible variations on each feature point). Extracted shape and texture features from these feature points are first projected into their own eigenspace using PCA. In eigenspace, the corresponding shape and texture weight vectors are further integrated through a normalization procedure to form an augmented vector which is used to represent each facial image. For a given test facial image, the best match in the model library is identified according to a similarity function. Experimental results involving 20 persons with different facial expressions and extracted from different view points have demonstrated the efficiency of our algorithm.