Intensity Inhomogeneity Image Segmentation Based on Active Contours Driven by Self-Similarity

Abstract

The segmentation of images with intensity inhomogeneity is always a challenging problem. For the segmentation of these kinds of images, traditional active contour models tend to reduce or correct the intensity inhomogeneity. In this chapter, we present a framework to make use of the intensity inhomogeneity in images to help to improve the segmentation performance. We use self-similarity measure to quantify the degree of the intensity inhomogeneity in images and incorporate it into a variational level set framework. The total energy functional of the proposed algorithm consists of three terms: a local region fitting term, an intensity inhomogeneity energy term, and a regularization term. The proposed model treats the intensity inhomogeneity in images as useful information rather than alleviates the effect of it. The proposed method is applied to segment various intensity inhomogeneous images with promising segmentation results. Comparison results also prove that the proposed method outperforms four state-of-the-art methods.