Recognition Authentication Technology

Since ancient times human authentication has gone through a very long period of change. With the advent of digital age, technologies based on biometric authentication have shown their value. Face detection technology is based on face features for identity recognition. Beacause face is always exposed, facial feature collection has a direct, friendly, convenient advantage compared with other means of identification. The automatic face recognition has become an important research point in the area of artificial intelligence and human biometrics and a very active subject in pattern recognition and computer vision.By constructing face detection module and face recognition module to a face recognition authentication system, an improved BP neural network and non-balanced binary tree support vector machine with clustering are proposed, and experimental results validate this method effective. Concrete work done as follows:First, a simple introduction of face recognition technology background and practical significance is presented as well as theoretical overview of the preliminary work.Second, based on performance of traditional support vector machine, non-balanced binary tree support vector machine is presented. By FastICA algorithm for image compression, to achieve the purpose of face detection though improved adding additional momentum, variable learning rate BP neural network face detection.Again, After Gabor filter feature extraction, Hopfield neural network compression, with non-balanced binary tree support vector machine, correctness and reliability can be valified though data analysis.Finally, the application to the real authentication process further validates this system by testing the integrity and availability. At last, a simple view of the prospects in this area is provided.