Local image segmentation model via Hellinger distance

Highly accurate active contour models are widely used in various image segmentation methods. In this paper, we propose an image segmentation model based on Hellinger distance for local region intensity fitting (HD-LRIF). The method defines two different metrics based on the Hellinger distance and constructs a new data fitting term to segment the image efficiently by minimizing the energy function. In addition, our method is independent of the initial contour and the segmentation results consistently obtain high accuracy. The experimental results show that the HD-LRIF model is far superior to state-of-the-art segmentation methods in terms of accuracy and efficiency. Specifically, it can effectively filter the noise interference, thus enhancing robustness and improving the accuracy of image segmentation in general.