Generalized Chan-Vese Model for Image Segmentation with Multiple Regions
Dang Tran Vu1, Tran Thi Thu Ha1, Min Gyu Song1, Jin Young Kim 1, Seung Ho Choi2, Asmatullah Chaudhry1, 3
1School of Electronics & Computer Engineering, Chonnam National University, 300 Yongbong Dong, Buk-gu, Gwangju, 500 757, South Korea
2Department of Computer Science, Donshin University, Information Center, Kunjae-ro Naju, Chonnam, 520 714, South Korea
3HRD, PINSTECH, P.O. Nilore, Islamabad, Pakistan
Abstract: In this paper, we propose a modified region-based active contour model by integrating the local information of foreground region into Chan-Vese model. Considering local spatial information term in the conventional energy function, our proposed model is able to overcome two limitations of the previous region-based level set methods: a) high sensitivity to the location of initial contours, 2) inability to segment multiple objects. Experimental results show the effectiveness of our proposed method as compared with other major level set-based techniques in terms of both efficiency and accuracy for 2D image segmentation.
[Vu DT, Ha TTT, Song MG, Kim JY, Choi SH, Chaudhry A. Generalized Chan-Vese Model for Image Segmentation with Multiple Regions. Life Sci J 2013;10(1):1889-1895] (ISSN:1097-8135).http://www.lifesciencesite.com.
Keywords: Generalized Chan-Vese model, region-based active contour, level set method, image segmentation, local information.