Face Recognition Using Local Gabor Phase Characteristics

This Paper proposes a new face recognition method based on local Gabor phase characteristics. In our proposed method, according to the good spatial position and orientation of Gabor filter, a Gabor filter with four frequencies and six orientations is firstly applied to filter face images. Based on daugman´´s method and the local XOR pattern, local Gabor phase patterns are then extracted to form the characteristic images. Finally, fisher linear discriminant analysis is used to project the characteristic images of each spatial position and orientation into low dimensional space. Nearest classifier is adopted to the projected characteristics to get the recognition result. Two human face databases, namely Feret and AR database are selected for evaluation. Experimental results show that our method consistently outperforms other recognition methods based on PCA, fisher linear discriminant analysis and Gabor magnitude characteristics.