ROI Extraction in Dermatosis Images Using a Method of Chan-Vese Segmentation Based on Saliency Detection

Abstract

Abstract. Extraction of ROI (Region-Of-Interest) in dermatosis images can be used in content-based image retrieval (CBIR). Image segmentation takes an important part in it. And the performance of the segmentation algorithm directly influences the efficiency of the ROI extraction results. In this paper, a method of Chan-Vese segmentation based on saliency detection to extract the ROI of the dermatosis images is proposed. Firstly the spectral residual approach (SR)[11] is used to get the saliency map of the dermatosis images. Secondly threshold segmentation is used to get the initial ROI images. Finally the Chan-Vese model is used to segment the initial ROI images to get the final ROI images, which can ensure the active contours evolve close to the object and remove the redundant information from the complex background. The experiment results show that the proposed method has the better performance than only using Chan-Vese method.

Cartoon-texture evolution for two-region image segmentation

“…The scheme presented in (15) can be further simplified by exploiting the linearity of the constraints H u − z = b , as suggested in [44]…”

“…Ω in and Ω out are the foreground and the background of the image, respectively. Although the choice of m = 2 significantly simplifies the segmentation problem, it has a lot of application fields, such as biological and medical imaging, text extraction, compression of screen content and mixed content documents, and can be used as a computational kernel for more complex segmentation tasks [15][16][17][18][19][20].…”

Cartoon-texture evolution for two-region image segmentation

“…and we used χ [0,1] (u) to indicate the characteristic function of the hypercube [0, 1] nx×ny . Consider the Lagrangian and the augmented Lagrangian functions associated with problem (11), defined respectively as…”

“…Theorem 1 Consider problem (11) where F (u) and G(z) are closed, proper and convex functions and H has full rank. Consider the summable sequences {ε k }, {ν k } ⊂ R + and let Cartoon-texture evolution for two-region image segmentation…”

“…Ω in and Ω out are the foreground and the background of the image, respectively. Although the choice of m = 2 significantly simplifies the segmentation problem, it has a lot of application fields, such as biological and medical imaging, text extraction, compression of screen content and mixed content documents, and can be used as a computational kernel for more complex segmentation tasks [11][12][13][14][15][16]. A widely-used two-region model was introduced by Chan and Vese in [17] and, together with its variations, is regarded as state of the art in the segmentation community.…”