Active contour driven by local entropy energy function for segmentation and bias correction
Gai Pan, Liqun Gao, Zhaohua Cui · 2013
C-V model has poor segmentation of images with intensity inhomogeneity. To overcome that problem, a novel active contour driven by a local entropy energy function is proposed to segment images with intensity inhomogeneity and get bias corrected images. The main idea of this paper is entropy can measure the degree of intensity homogeneity and local intensity information is extracted due to a weight function. Simulation experiments of 3 images show: this method can deal with intensity inhomogeneity problem of C-V model, and has better adaptability to initial location of the contour curve.