A novel active contour for medical image segmentation

Abstract

In this paper, a novel parametric active contour is proposed for image segmentation. It integrates the wavelet transform and self-affine mapping system to effectively compute the external force of parametric active contours. For this purpose, after smoothing the image by a Gaussian kernel, the wavelet coefficients are computed. Then, sub-forces of each wavelet scale were computed by using contractive self-affine maps. Finally, the total self-affine force field was obtained by superposition of these sub-forces. Experimental results demonstrated the superior performance of the proposed algorithm compared to a number of frequently-used active contours for medical images.