GENERALIZED CHAN-VESE MODEL FOR IMAGE SEGMENTATION WITH MULTIPLE REGIONS

摘要

In this paper we propose a modified region-basedrnactive contour model by integrating the local informationrnof foreground region into Chan-Vese model. By adding arnlocal fitting term into the conventional energy function,rnour model is able to overcome two limitations of thernprevious region-based level set methods: high sensitivityrnto the location of initial contours and inability to segmentrnmultiple objects. We will show the advantages of ourrnmethod compared with other major level set-basedrntechniques in terms of both efficiency and accuracy byrnsegmentation experiments.