A Mixed Face Recognition Method Based on ID Card
Zhigang Chen · Jisuanji gongcheng · 2003
Based on the analysis of human face in ID card, this paper proposes a new face recognition method, which integrates the geometric eigenvector matching with the weighted face subdivision matching. The current recognition methods cannot effectively smooth out the disturbance caused by all kinds of face expressions, so the presented method firstly makes a rapid and correct location of the eyes, balances and normalizes the image. Then on one hand, it extracts the geometric eigenvectors, which can avoid the expression disturbances; on the other hand, it makes a weighted face subdivision matching on the normalized faces. Finally it outputs the final result based on the two previous matchings. The experimental results on the JAFFE face database show that the given method not only has a high recognition rate, but also can effectively smooth out the face expression disturbances, so it gets a satisfactory result.