Face Recognition Based On Gabor Wavelet And Local Binary Coding

Face recognition is one of the major hotspots content and pattern recognition and image processing research, its research began in the late sixties of the twentieth century, as a category in the field of biometric identification technology based on physiological characteristics, it is through the computer extract facial features of people, and according to the extracted features for authentication, in the past 40 years has been considerable development. In recent years, information processing and transmission means based on information society, it is quietly changing the lives and the way the management of human society, whether it is to improve the condition of the hardware, or the actual needs, face recognition technology has gradually become an urgent need hotspot.This Paper from the facial feature extraction algorithm start by studying classical face recognition algorithm based on Gabor wavelet feature extraction and recognition Local Binary coding algorithms, analyze their advantages and disadvantages, and the corresponding improved algorithms. The main work is as follows:(1) The study based on Gabor wavelet facial feature extraction algorithm, due to the classical Gabor wavelet facial feature extraction algorithm feature high dimensionality, time consumption, this paper presents an improved algorithm, the first analysis of different factors and different scales The Gabor wavelet recognition rate factor, to elect a set of "best" Gabor wavelet to extract facial features, and the face image is divided into 8 nonoverlapping sub-image is calculated and given different weights for each sub-image performance value, and experimental verification algorithm.(2) The face recognition algorithm based on local binary coding techniques. Analyze the advantages and disadvantages of the classic local binary encoding techniques, compare their improved operator annular symmetrical local binary encoding technology(CS-LBP) and multi-level regional binary encoding mode(MB-LBP) and the traditional LBP operator child's performance, and a multi-stage MB-LBP face recognition algorithm, compared with the traditional LBP operator, multistage MB-LBP face recognition algorithm not only effectively improve the recognition rate algorithms, but also effectively improve the algorithm robustness.(3) Extract facial features and advantages of the LBP operator for Gabor wavelets, an adaptive face recognition algorithm based on Gabor wavelet and CS-LBP's(ASLGBP), characterized by adaptive weighted based on the integration of the sub-image Gaborpeople Face recognition algorithm experiments we know the face sub-image research, nose, mouth and other facial region more details in recognition heavily weighted, so ASLGBP Firstly, by integrating Face projector will face the eyes, nose, mouth et al face area intercepted, then followed with Gabor wavelets and CS-LBP extract facial features, experimental results show that compared with some classical face recognition algorithm, ASLGBP algorithm effectively improve the recognition rate.