An unconstrained hybrid active contour model for image segmentation

In this paper, we propose an unconstrained active contour model combining edge and region information for image segmentation. The new method achieves the segmentation by alternating the regularization term and the data-fidelity term. We use a morphological approach to the regularization term which is the most time-consuming in the energy function. The proposed method is robust to noise and avoids re-initialization. The efficiency of our method is validated by testing it on various images.