3D Face Recognition Based On Contour Lines And Geometric Feature Vector

Compared with2D face recognition,3D face not only contains richer information but also canovercome the effects of pose, illumination, expression, makeup etc. Because it can adequatelyrepresent the human face, it is an efficient way to execute3D face recognition which includesacquisition and pre-processing of3D data, feature extraction and design of classifier.To resolve the keys of3D face recognition, the research work of this paper is as follows:(1)Data preprocessing. Locate the nose and rotate the face to positive direction based on PCA(Principal Component Analysis). And then it is needed to cut the face according to Geodesicdistance. To reduce calculation, we need double-cubic B-Spline surface fitting.(2)Contour lines and geometric feature vector extraction. We extract the carve contour andhorizontal contour based on curvature. And then, we locate the key feature points of face on searchmethod of near area.(3)3D face recognition system. ICP (Iterative Closet Point) algorithm and the average Hausdorffdistance are discussed in this paper. Three classifiers are designed.(4)Experimental results. In GavabDB database, different methods of feature vector extractionare applied to compare the recognition rate. We identify the effect of using different classifiermethod. The experimental demonstrates that using method in this paper can get92%.