Face Recognition Technology Used In Digital Video Surveillance System

Face recognition is one of active research topics in the fields of pattern recognition and machine vision. It is also a challenging problem. Digital video record system is a widely used product recently. It is a development direction of digital video record system that face recognition is used in it. In this paper, we will discuss what we have done on this problem.The main contributions of this paper are as follows.1. Some essential knowledge of images and some means of dealing with images are introduced, which is often used in face detection and face recognition.2. Methods of Motion detection, skin color detection, Singular Value Decomposition(SVD) and Support Vector Machine(SVM) are introduced. These methods are often adopted in face detection. Advantage and disadvantage of these methods in the condition of using in digital video record system are discussed.3. Face recognition methods of Karhunan-Loeve(KL) transform and neural network are discussed. Advantage and disadvantage of these methods are analyzed.4. A new method based on combination of local SVD and SVM that is used in face detection is presented. Local SVD is used in this method, and local SVD vector is used as input vector of SVM. This reduces computation and accelerates face detection. And this can also uses information of images more sufficiently. This makes detection more accuracy.5. A new method based on combination of local SVD and neural network that is used in face recognition is presented. This method has the advantage of low computation, high accuracy and strong ability of anti-jamming.6. A new method is presented which can detect face and recognize face at the same time. The advantage and disadvantage of this method is discussed. The approach of recognition is simplified, and the speed is improved. Experiment is done and the result is analyzed.A recognition system is built and experiments are done. The theories and methods we present here is available and practical.