基于Gabor小波和LBP算子的人脸识别技术研究-微电子学与固体电子学专业论文.docxVIP

RESEARCH ON FACE RECOGNITION METHOD USING GABOR WAVELET AND LOCAL BINARY PATTERN ABSTRACT Face recognition is currently one of the hottest research areas, it is well worthy of our study either in theory or in practical application. The choice of feature extraction methods and their performance will have a direct impact on the whole system of face recognition. Our work is focused on some issues related to feature extraction methods for face recognition. The emphases of our work is the feature extraction algorithm based on gabor wavelet and local binary pattern with its application to face recognition. The work and the innovation in this dissertation can be summarized as following. Research on face recognition method using Gabor wavelet and linear discriminnant analysis. This method makes full use of the desirable characteristics of spatial locality and orientation selectivity of Gabor filters. First, Gabor features are extracted on the original intensity images, then the best classification feature is acquired by utilizing the LDA. Special nearest neighbor classifier based on face feature matris is used for classification. Take experimentation on ORL and YALE. Research on method—Gabor wavelet and locality preserving projections. Research on face recognition method using Gabor wavelet and locality preserving projections. Locality preserving projections algorithm takes into account the space structure of the samples, and in the process of dimension reduction, it can thus find a good linear embedding that preserves local structural information and intrisinc geometry of the data space. Take experimentation on ORL and YALE. Research on feature extraction method using Gabor wavelet and locality preserving projections. Improvements of this paper is that: after extracting the histogram features by combinating the Gabor wavelets and the LBP, used the LPP algorithm reduce its dimension further, which can extract the most beneficial features for the classification while reduci