I am using Chan Vese active contour to segment some Ultrasound images. I was wondering whether there are several ways to make this contour more robust or somehow to define which of the grey tones it should segment, rather than binarize the images.
I know that it utilizes an energy minimization function based on a Mumford Shah technique. I guess there must be a way to tune some of its regulatory parameters and train the contour to segment specific tones. As I said again, I don't want to threshold the images. I want to modify the contour's behavior.
Any ideas please?