Research On Face Recognition Systems Under Complicated Circumstances

With the development of statistics, informatics, computer science, etc., the technique of face recognition has gained a great development. As a main technical means of biological discrimination, face recognition has been widely used in the fields of video surveillance, information security, image retrieval, ambient intelligence and so on. Because of environmental disturbance, limitation of shooting conditions, complexity of faces, etc., the face recognition rate can’t meet the requirement of practical applications. Meanwhile, the complexity of face recognition algorithm will influence the real-time performance of face recognition systems.So, there is still a long way to go to the actual applications of the existing face recognition systems.Aiming at the problems of face recognition, a series of researches have been carried out in this thesis. Compressed sensing arose in recent years, which is a new theory of signal processing. This thesis has probed deeply into the applications of this theory on face recognition.The main works done in this thesis are as follows:1. On the basis of deep research on compressed sensing(CS) theory, we have proposed a face recognition algorithm based on CS.2. We’ve presented an optimization design method for measurement matrix which corresponds to the theory of CS. And we then prove the feasibility of this algorithm by experiments of face recognition.3. We have designed and realized a face recognition platform based on MATLAB. The procedure includes face recognition systems designment, systems’ modules designment and the realization of face recognition platform by MATLAB. We’ve conducted experiments about the influence of di?erent factors, such as preprocessing, number of training samples and dimensions of extracted features, to face recognition rate upon the MATLAB platform, and then analyzed the results of the experiments.4. In order to solve the problem of low face recognition rate under the complicated circumstances, we can improve the recognition rate by face detection and extraction. With the function of face detection through OPENCV, we’ve designed a face recognition platform based on VC++. So we can recognize the faces under complicated circumstances by this platform.