Active contour driven by region-scalable fitting and local Bhattacharyya distance energies for ultrasound image segmentation

In this study, a new local region-based active contour model in a variational level set formulation for ultrasound (US) image segmentation is proposed. An energy function based on the region-scalable fitting (RSF) term and the local Bhattacharyya distance term is formulated. The RSF term is dominant near object boundaries and responsible for attracting the contour towards object boundaries, and the local Bhattacharyya distance term including local region statistical information improves the robustness of the proposed method. The model can handle blurry boundaries and noise problems. In addition, the regularity of a level set function is intrinsically preserved by the level set regularisation term to ensure accurate computation. A level set function is used to define the partition of image domain into two disjoined regions. Experimental results demonstrate the desirable performance of the proposed method for synthetic images with different levels of noise and US images with weak boundaries and noise.