A Segmentation Algorithm for Medical Images Based on Level Sets
Tang Qingy · Beijing shengwu yixue gongcheng · 2005
The active contour method proposed by Chan and Vese is improved. The boundary statistics is used instead of the region statistics in Chan-Vese's method. In every iteration of image segmentation, firstly the corresponding boundary statistics of the current curve is found, then it is driven in evolutive equation, finaly the curveis moved to capture the object's boundary under the level sets framework. Experimental result shows the effectiveness of the proposed method.