Abstract:

The gradient level set model has several disadvantages, it is sensitive to noise, it is unsatisfied on keeping the image edge, the segmentation result depends on initially parameters, and the segmentation process can not stop when obtaining the optimal solution. In order to solve the problems, a level set image segmentation algorithm based on cultural algorithm is proposed and the cultural algorithm is applied to the C-V (Chan-Vese) level set model. Firstly the parameter selection is automatically realized. Secondly the situational knowledge and the normative knowledge are used to guide the population evolution in belief space. And finally the image segmentation process is timely stopped by judging a change in the image entropy fitness value. The experimental results show that the algorithm is superior to conventional methods in the anti-noise performance and segmentation efficiency, and can accurately segment the medical image lesion areas.