The Research On Method Of 3D Face Recognition Based On Binocular Stereo Vision

The thesis focuses on 3D face recognition based on binocular stereo vision problem mainly.Based on Marr vision theory,it has focused on the research of binocular stereo vision of camera calibration,edge detection,stereo match and existing 3D face recognition algorithms.In camera calibration,Tsai's method is adopted to get the exterior and inner parameter.; For edge detection,firstly do median filter to image in order to wipe off noise,the make use of Gauss-Laplace operator to pick up edge;For stereo matching,a feature-based and phase-based method is presented.This method makes good use of the respective advantage of feature-based arithmetic and phase-based arithmetic;In the part of "3D face recognition", existing recognition algorithms are carefully reviewed and discussed in theory.The approaches including PCA,LDA,ICP,and Hausdorff are validated and analyzed with experimental results.