Near-infrared Face Recognition System Based On Dm6446

Face recognition technology is adopted widely as one of all biometrics identification technology. The conventional personal identification methods such as identity card, password and key are not satisfied to the need of modern society in security and convenience. The invariability and solitary of face recognition produce reliable warranty of face recognition as a mean for identification. Automatic face recogniton is becoming a fact with the development of computer image processing and pattern recognition technology.Most of traditional automatic face recognition system is based on PC online application. Under the driving of information security and technology evolution, face recognition is applied widely in criminal identification, business management and self-service, it is required to be easier to carry and use. This thesis research and development an automatic face recognition system based on DSP.Firstly, application trend and existing problems of face recognition are investigated based on much of what other people have done about face recognition technology. The solution is proposed based on davinci technology.Secondly, an automatic face recognition system based on DSP is put forword. Six module including power, processor, image acquisition, memory, man-machine and peripheral are designed detailedly.Thirdly, a set of algorithm which is quite suitable to this hardware system is proposed, three stages of the whole algorithm including image preprocessing, Adboost and 2D-PCA are discussed and optimized in detail.Finally,analysising the parameters which influenceing the system performance using the ORL database. Mass experiments indicate that this face recognition system has following features: low power, small size, high speed, high accuracy and steady operation, meeting the requirments of real-time face recognition.