The Research And Implementation Of Face Recognition Algorithm Based On Features Fusion

In recent years, antiterrorism, homeland security, and social security are attracting more and more attention. Identity authentication is the core of solving these problems and the recognition accuracy, safety and real-time is becoming more and more important. Among the biological feature recognition, human face recognition occupies an important position. Compared with other biometric features recognition(fingerprint identification, iris recognition, etc.), face recognition is optional, non-contact and concealed. Face recognition is widely used in the field of security monitoring, authentication, image-based search and human computer interaction(face unlock, face electronic payment, etc.).This paper analyzed the core problem and difficult points of face recognition technology and focused on weakening the effects of the illumination and expressions. Face recognition technology mainly contains five modules: facial image collection, face region detection, image preprocessing, feature extraction and feature classification, etc. First, this paper introduced and researched several processes of face recognition technology in details, including: face detection, image preprocessing, face image feature extraction and classification, and implemented some key algorithms. Second, this paper conducted a lot of simulation experiments based on the principal components analysis(PCA and 2DPCA) and analyzed theoretically the advantages and disadvantages of the methods. Third, this paper proposed a hybrid feature method — based on principal component analysis and rotation invariant uniform local binary pattern texture features — in order to weaken the effects of varying illumination conditions and face expressions. On the PC platform(MATLAB2011,VS2010,Open CV), this paper carried on a lot of experiments on the standard face image databases, random face image databases and homemade face database to evaluate the performance of the proposed algorithm. The evaluation results verified the superiority of the algorithm.In the process of hardware implementation of the algorithm, this paper realized the face real-time detection and identification based on PC. Then according to the characteristics of the selected DSP chip, the part of the algorithm was transplanted and optimized.