New Image Segmentation Models

T.D. Bui, S.Gao, and Q.H. Zhang (Canada)


Image segmentation, Mumford-Shah model, Level set method, Modified Mumford-Shah model, Constant and lin ear approximations.


In this paper the Mumford-Shah (MS) model and its vari ations are studied for image segmentation. It is found that many previous models cannot detect edges with low con trast. We have studied a variety of different models of energy minimization a la Mumford-Shah approach for im age segmentation. All these new models are better than the piecewise constant approximation since they can detect the low contrast edges of objects. This paper will present and compare these models for segmentations of medical and natural images.

Important Links:

Go Back