基于水平集方法的图像分割研究-模式识别与智能系统专业论文.docxVIP

燕山大学工学硕士学位论文Abstract 燕山大学工学硕士学位论文 Abstract Image segmentation is one of the key problems in computer vision． Classical image segmentation techniques cannot satisfy the requirements of complex image segmentation due to the limitations itseIf；Under the circumstances,modebdriven image segmentation techniques come into beiIlg and become a focus of widespread concerto Geometric Active Contour Model is capable ofhandling changes in the topology ofthe evolving contour,it is the difficult problem to Parameter Active Contour Mode．And the emergence ofthe level set method,which has greatly promoted the development ofthe geometric active contour models． This paper mainly based on the level set method for image segmentation． In reviewing the development ofdigital image processing，image segmentation process and the current situation based on the researc]z,theory of Active Contour Models，cnrve evolution and level set method are presented． Traditional level set methods rely on computing the evolution of all the level sets．Using particle swami optimization and Barrow band，Parameter Active Contour model combined with level set，thereby reducing the amount of calculation to save calculation time；Aiming at C-V method based on simplified Mumford-Shah model,proposed algorithm and the corresponding iterative termination criterion,which can improve the speed and reduce the number of itemtions；Narrow band method is not used directly in C-V method based on simplified Mumford-Shah model，an segmentation method ofnarrow band and Mumford-Shah model is presented，calculation is limited in the range ofnarrow band,SO the computational complexity decrease；Tbe velocity field for the moviIlg curves generated by this elastic interaction between image boundaries and the moving curves，this paper proposed a edge-based level set algorithm based on new velocity field,level set method is used for the calculation of Ⅱ Abstractiteration,a Abstract iteration,a length scale is obtained by using nonlinear diffusi