A New 3d Face Recognition Method Based on Combined Category Information

In view of the shortcomings of existing face recognition that there are changes in face pose or changes in light intensity and angle, it is impossible to make full use of the training sample category information for training, resulting in recognition errors. The image feature space decomposition and extraction in this paper adopts the combined category information. The non-negative matrix factorization method can make full use of the category information of training samples and improve the recognition performance. The experimental results show that compared with the traditional non-negative matrix factorization method, the method adopted in this paper has better improvement and promotion in the effectiveness of feature extraction and face recognition rate.