Region-edge-based active contours driven by hybrid and local fuzzy region-based energy for image segmentation

This paper raises a region-edge-based active contour driven by the hybrid and local fuzzy region-based energy to segment images with high noise and intensity inhomogeneity. The energy functional consists of region energy and edge energy. The region energy is made of hybrid fuzzy region term and local fuzzy region term. Its aim is to motivate initial contour to move toward the exact object boundary. What's more, it is proved to be convex and ensures the segmentation results independent of initialization. The hybrid fuzzy region term can balance the importance of the object and background while the local fuzzy region term by incorporating spatial and local information can decrease the effect of intensity inhomogeneity in given images. The edge energy is used to regularize the pseudo level set function (LSF) and maintain the appearance of the smoothness during the curve evolution. Inspired by the fuzzy energy-based active contour (FEAC), a more direct and simpler method is developed to calculate the difference between the old and new energy functions to update the pseudo LSF during the curve evolution. Experimental results on synthetic and real images with high noise and intensity inhomogeneity show that the proposed model can obtain better performance than the state-of-the-art active contour models. The code is available at: https://github.com/fangchj2002/HLFRA. (c) 2020 Elsevier Inc. All rights reserved.

Automated summary

This paper presents a novel active contour segmentation method with region-edge driven by hybrid and local fuzzy region-based energy. The method achieves good segmentation results for images with high noise and intensity inhomogeneity and demonstrates better performance than other methods in experiments.