Study Of Video-based Face Segmentation And Recognition Techniques

With the development of computer network and information technology, an accurate automatic personal identification is increasingly critical. International academic circles and some enterprises pay close attention to this research area. Biometrics refers to automatic identification technology of an individual based on their physiological traits such as fingerprint, face and iris or behavioral traits such as signature, speech and gait. Facial images are probably the most common biometric characteristic for personal identification. Face recognition is a non-intrusive technique and people generally do not have any problem in accepting face as a biometric characteristic. So it can be used in various fields, such as security control, criminal capture, access control, electronic commerce, welfare benefits disbursement, video supervision and videoconference. Face detection and recognition is one of the hot issues in pattern recognition and artificial intelligence fields, which involves image processing, psychology, physiology, recognition science and so on, so the research on this issue is challenging and meaningful.On the basis of analyzing and summarizing the relative research achievement and the newest development, this thesis make systemic experiments and theory research on face detection and face recognition, some achievements are obtained in video-based face detection and recognition technology.Main contributions of this thesis include face detection in video images and face recognition based on lifting wavelet transform and Learning Vector Quantization (LVQ) neural network. We try to integrate the technique of Markov Random Field, arithmetic morphology, machine learning, optimization theory, wavelet transform, and neural network with the theory of intelligent image processing in computer vision, and puts forward and realizes a set of methods and algorithms that are valuable for face detection and recognition system.Automatic detection and localization of human faces are difficult tasks. In this thesis, we propose a face location method that integrates the moving information in video...