关键词: 相位一致性特征, log-gabor滤波, 分块, 主元分析, 人脸识别

Abstract: The recognition accuracy of visible images is variant to changes in illumination conditions and facial expressions, and variations in face images are confined to local regions, while the phase congruency technique used in this research employs the local energy model, which is invariant to changes in illumination and contrast conditions in images. A method of face recognition based on modular phase congruency is used in this paper. The face images are convolved with a bank of 2D logarithmic gabor filters, and the images of phase congruency are obtained by using the model of phase congruency; the extracted feature maps are dealt by modular PCA; considering all the sub-images distance information, the nearest distance classification is used to distinguish each image. Experimental results show that the proposed method has better recognition performance.

Key words: phase congruency, logarithmic gabor, modular, Principal Component Analysis（PCA）, face recognition