Face Recognition Program for Person Identification

Trends in computer vision and pattern recognition and capabilities of modern computers contributed to a considerable amount of researches of these areas application in facial recognition systems. The purpose of this study is to investigate the most significant appearance-based and local features based methods of face recognition, to choose the most relevant one and to implement the face recognition system built on top of the chosen method. In the first part of current research methods of face recognition and identification are presented. The analysis of methods covers the most important features of pattern recognition by examining algorithm structure. An application of groups of methods is considered for different purposes. Second part reveals the result of algorithms testing and details of face recognition program implementation. Current study contains thorough comments for capabilities of each algorithm under observation, which are verified in the second part by testing algorithms using real-world datasets and examples. On the basis of the results of this research, the program of person identification is implemented and tested, it uses a SVM classifier for person identification, which proved itself to provide a 96% precision rate of a person identification after fourth training session. Current research may be cognitive for researchers in the field of facial recognition, it contains a concise information about various methods and their applications in pattern recognition area.