A novel active contour model based on features for image segmentation

Abstract

Active contour model is an extraordinarily valuable technique in image segmentation, which is essential for image analysis and understanding. Active contour model has been widely studied because it delineates closed and smooth contours or surfaces of target objects. However, traditional active contour models underperform on complex natural images. To tackle this problem, we propose a novel active contour model framework, called FeaACM. We introduce the feature energy function into the conventional energy functional to minimize the energy functional to maintain the consistency of the object region and account for different distributions of objects and backgrounds in the feature space. To demonstrate the advantages of our method, we compare our method with the state-of-the-art methods, and show that our method achieves competitive performance. In addition, we utilize AutoEncoder technology to extract the feature of the image verifying the generality of our framework. Extensive and numerous experiments indicate that our method can segment complex natural images effectively. Our code is available at https://github.com/xuepeng1234/FeaACM.