Based On The Three-dimensional Reconstruction Of Face Recognition

Biometric recognition technology, which identifies people through their unique features, becomes hot recently due to its convenience and efficiency. As we know, human's many biometric features, such as finger-print, iris and voice could work only when human take initiation with the machine, but face recognition is not in this limitation and is natural, harmonious, undisturbed and easily accepted by users. Face recognition, with so many advantages, is attracting more and more attention.In modern time, as personal identification technology is applied almost everywhere, the biometric recognition based on finger-print, iris, face and so on recognition technology can be needed in many occasions, such as guarding system at roll-gate, check in system in airdrome, monitoring system used as automatically registering, validating, scouting in the electron business affairs between natural person and machine and other monitoring fields.Based on digital image procession technology and graph image relative theories, and combined computer vision technology with mode recognition technology, the paper deeply researches and analyses the face recognition system. And we have finished following works in this article:(1) Detection of human face. Two face location methods are put forward in this article,which can locate human face region exactly from the human images. The first one is face location based on skin color and face template matching, and the other method is based on the ASM algorithm detecting the human face outline.(2) Reconstruction of special human face. One algorithm based on Candide-3 standard model is adopted in the article. We can get the special person's reconstructed face by adjusting the parameters of Candide-3 model to the data of face image.(3) Recognition of human face. Two recognition algorithms, based on the Hausdorff Distance algorithm and based on the Gabor function, are discussed in the part. By comparing the results of the two algorithms the paper put forward one new recognition algorithm of human face. The experiment result shows the new algorithm can be feasible.