基于透视原理的人脸三维重构研究-计算机应用技术专业论文.docxVIP

基于透视原理的人脸三维重构研究ABSTRACT 基于透视原理的人脸三维重构研究 ABSTRACT Now,face recognition is mainly based on 2-D images．But much solid information is lost in the 2-D image．If 3-D face is used in the technology of face recognition,the results of face recognition will be more correct，the applied areas will be more extensive,and the meanings will be more important． 3一D face recognition refers to the technology of 3-D reconstruction,and its basic materials are 2-D images．We study the technology of 3-D face reconstruction based on perspective theory,and the foundation of the next study of 3-D face recognition in the future iS established．In the research of 3一D face reconstruction based on perspective theory, we detect the face from complex background,theposition of the face is determined,and then feature points are extracted firstly．Secondly,camera calibration is carried on￡so the parameters of the canlera are obtained．Thirdly,we study the the research of 3-D face reconstruction using feature points and camera parameters． In the face detection,skin-color model ofthe face is established，and we carry out the face detection based skin-color by using Support Vector Maehine(SVM)．First，we select the appropriate color space model,the pixel’S color information Call be taken as the main feature．And Sequential Minimal Optimization aigortthm is used to train the SVM in order to segment chromatic image．Then we study face location．We search the border ofthe face by horizontal and perpendicular projection．A rectangle is drawn at the border ofthe face． By tlliS method，the detection effect is good under the standard lighting effects．Face feature is matched after face detection．In this paper,we introduce three matching methods： correlation method，templet method and the matching method based on face feature．We analyse the advantage and disadvantage of the two methods above．And we use the matching method based on face feature．It is regarded as the improved one of the two methods above．The horizontal and per