Hybrid active contours in multiphase level set framework for images segmentation

A major problem with image segmentation is the building of model that is able to deal with all kind of image. This is due to the diversity of the image sources. However, the aim is to widen, as much as possible, the capability of the model to segment several image modalities. Hybridization between some models seems a good alternative to achieve that. In this paper, functional that incorporate several kinds of image information is used: edge detector function, local means and variances, and global means. Such choice allows getting successful segmentation results as it will be shown in the experimental section.