Recently in the study of medical image segmentation, see the level of concentration of chan - vese model, using the found many problems, don't know how to solve, to ask questions,

Level set in common image segmentation effect is very good, probably should be the difference between foreground and background is large, such as the following figure


But in liver image segmentation, target, and there are some adhesion, segmentation effect is very poor, see the level set initialized contour and evolution of others can split out what you want in the multi-objective area, I also try to initialize the different profile, but no effect ah,


Did the level set algorithm has a very good understanding of the teacher teach me how to optimize or is how to use the level set algorithm in the segmentation of liver CT images?

CodePudding user response:

How do you define the initial contour?