Study Of Face Recognition Technology

Face recognition has very large academic and practical values. In daily life, people knowing each other uses at most of person's face. Face is the most familiar model in human vision. The visual information reflected by face has important meaning and impact between people's intercommunion and intercourse. Because of its extensive and applied realm, face recognition technique has got the extensive concern with study in near three decades and become the most potential method of identity recognition. At the same time, it is difficult to implement face recognition using computers. First, human face is a deformable object composed of complex 3D curve surfaces, which is hard to be represented in form of mathematics. Secondly, faces of different persons have the similar structure, and the face images are greatly dependent on ages and photography conditions. This paper mainly study face extraction and class method, which concept can be summarized as follows.In this paper, Independent Component Analysis (ICA) is presented as an efficient face feature extraction method. In a task such as face recognition, important information may be contained in the high-order relationship among pixels. ICA is sensitive to high-order statistic in the data and finds not-necessarily orthogonal bases, so it may better identify and reconstruct high-dimensional face image data than Principle Component Analysis (PCA). Conventional ICA algorithms, such as Informax algorithm, are time-consuming and sometimes converge difficultly. Informax algorithm need people adjust learning speed. In this paper, a modified FastICA algorithm is developed, which only need to compute Jacobian Matrix one time in many iterations and achieves the corresponding recognition effect of FastICA. Class descriminability select optimal independent components (ICs). The experiment results show that modified FastICA algorithm quickens convergence and ICs selection optimizes recognition performance. ICA based features extraction method is robust to variations and promising for face recognition.In order to better integrate face features for efficient classification, Hidden Markov Models (HMM) are used as to construct face models in this paper. Because HMM can keep the states unchanged for a given range of the change of observation vector and HMM use face similar construction. By analysis of HMM structure, tranditional 1D Hidden Markov Models (1D-HMM) has strongpoint of simple structure, Pseudo 2D HMM (P2D-HMM) is able to better model 2D data such as face images, due to its pseudo 2D structure. Combining these two models' defects and merits, an integrated model using Support Vector Machines (SVM) and HMM was proposed. Through Independent Component Analysis (ICA), some face area features are extracted. These feature vectors are as the inputs of SVM. SVM/HMM face recognition method achieved corresponding recognition performance, compared with P2D-HMM method, and experiment results show that SVM/HMM model has a simpler structure and lower computing complexity.Because face image is liable to impact of varieties and face is nonrigid and similar. Accurate face recognition is still difficult. There is still lone distance between face recognition and practicality. The progress of computer technology, pattern recognition, human intelligent and biologic psychology, vision mechanism surely promote face recognition develop.