Level set method by using local entropy and enhanced LoG operator for image segmentation
Feng Ding, Bin Ji · 2022 Global Conference on Robotics, Artificial Intelligence and Information Technology (GCRAIT) · 2022
Level set evolution is an essential method for image segmentation. However, the region-based level set is sensitive to contour initialization, and it is difficult to obtain good results in images with intensity inhomogeneity. We propose an improved level set method based on local entropy and enhanced Laplacian of Gaussian (LoG) operator for image segmentation. Specifically, the local entropy reflects the local intensity variations, which improves the method's ability to handle inhomogeneity. Then, the enhanced LoG operator smoothes the heterogeneous regions to reduce the influence of the regions on the curve. Experimental results show that the method can effectively segment inhomogeneous images and has good adaptability to contour initialization and noise.