The Research Of Infrared Face Recognition Based On Gabor Wavelets Transform

In recent years, infrared face recognition is a new hot spot of the research of biometric technology since infrared face images have lots of special properties such as defense from camouflage, cheat and independence to the ambient light. Infrared face recognition is a technology of feature extraction and classification on infrared images which are collected by thermal infrared imaging device and processed using the methods of pattern recognition and image processing etc. With the increasing development of infrared face recognition technology, it will have a great application value in the direction of security monitoring, access control and data confidentiality.In this thesis, the method of infrared face recognition based on Gabor wavelet is deeply studied based on the analysis of properties of the infrared image and the system flow of the infrared face recognition. The technology of feature extraction and classification of infrared face image is studied deeply from both feature extraction and classifier design.Firstly, the main processes of infrared face recognition system are discussed in this thesis. Then preprocessing technology of infrared face image is studied and analyzed deeply, which is very important for the process of the feature extraction and classification of infrared face recognition. The preprocessing includes angle correction and size correction. Preprocessed image is clear and has a high quality, which has laid a good foundation for the later accurately identification;Secondly, according to the imaging properties of infrared face image, including lower image resolution and blurred details, the method of feature extraction for infrared face recognition base on Gabor wavelet is studied deeply. And on this basis, a new method which combines local feature and global algebraic feature based on Gabor wavelet and SVD is proposed for the feature extraction of infrared face image. The experiments results show that, compared to traditional identification methods, this method has good recognition effect.Finally, the basic principles of ensemble learning are discussed. And on the basis of the previous work, a new method to construct and integrate multiple classifiers for infrared face recognition based on Gabor feature. Actually, this method is the use of multiple classifiers to generate complementary, making the total classifier to achieve the best classification results. The experiments results show that, the new proposed method has good recdgnition effect.