Feature extraction using Gabor feature-based IFDA

In this paper, we propose a new method of Gabor feature-based inverse Fisher discriminant analysis for face recognition. In the proposed method, the intrinsic feature is first characterized using Gabor wavelet transform with different scales and directions. Then, image discriminant features are extracted by selecting principal components and inverse Fisher disciminant vectors. Experimental results on ORL and FERET face database demonstrate the effectiveness of the proposed method.