The Research Of Face Recognition Based On Sparse Representation

The focus of this article is the face recognition. The feature extraction and the face classification recognition is also very important. If we have a given test face image, we can confirm the identified of image by using the training sample database. Most of the training samples need large image preprocessing and feature extraction. The face feature extraction has an effect on the face recognition rate, but it lacks robust about occlusion and corruption.The work of this thesis mainly focuses on face feature extraction and face classification. Based on the research of the two parts, compressive sensing and sparse representation algorithm for face recognition is put forward.Firstly, this article gives out a review on research background and developing roadmap of face recognition. Some kinds of face recognition algorithms were introduced and some conclusions were presented.Some research of Compressive Sensing(CS) are introduced, there are some key points need to pay attention when using CS for face recognition. The over complete dictionary is formed by using the training samples. A lot of research about the theory of compressive sensing and sparse representation have been used for face recognition, so it indicates that sparse representation has natural discrimination, it chooses the most compact subsets of the signal representation.We use the compressive sensing to reduce the dimension of face images, the sparse representation is employed to determine the class of the test face image. Experiment results show the CS theory and the sparse representation for face recognition has high accuracy and efficiency.Above the research on face recognition, we introduce the image segmentation and the face recognition algorithm we adapted for web-based face recognition system, the design and the realization of the system were also introduced.We do the experiments on the ORL and Yale database. The results show that the proposed algorithm has better performance in the feature extraction and recognition.At last, the article summarized the research work and pointed out the many direction of the future work.