Feature Extraction And Matching Of The Face Image Based On Wave

Computer binocular stereo vision is a hot research topic in stereo vision, the three-dimensional structural information of the objects can be restore by the images of an object taking from different angles, the data of three-dimensional structural information is the basis of three-dimensional object reconstruction. The problem from face images to obtain three-dimensional information of the face is researched in the paper based on computer binocular stereo vision technology, and the problem is related to image preprocessing, feature point extraction, feature matching technology.The main research works are as follows:1、Image preprocessing technology is researched, including gray processing, image denoising and image sharpen enhancement technology. Histogram equalization method is used in image gray processing; The neighborhood average method and wavelet threshold denoising algorithm are used in image denoising, wavelet threshold denoising algorithm is further researched, there is an analysis experiment about the Visushrink, Bayesshrink and SUREshrink threshold denoising method; For blurred image,laplacian masking method is used in image sharpen enhancement.2、The feature points are extraced based on the image edge. A new algorithm for edge detection is proposed, namely, firstly there is multi-scale decomposition for source image by column and row transform of self-adaptive lifting wavelet, secondly the wavelet modulus maximum algorithm and the Canny operator are respectively used to detect the edge information of high-frequency and low-frequency, then the edge of image is detected. There is comparative analysis about Canny operator, multi-scale edge extraction algorithm based on lifting wavelet transform, the wave testing method, the traditional wavelet decomposition method and edge detection algorithm proposed in this paper, the edge of face image is detected by the algorithm proposed in this paper, and feature points of face image are extracted by wavelet method.3、The feature of image is matched. Feature matching of face image is researched and analysed, wavelet pyramidal feature point matching method is used to match feature of face image through choosing appropriate measurement function of matching and layered matching strategy, then the accurate face matching rendering is made.