Entropy-based Kernel Graph Cut with Weighted K-Means for Textural Image Region Segmentation

Abstract

Recently, image segmentation based on graph cut methods has shown impressive performance on a set of image data. Although the kernel graph cut method provides good performance, its performance is highly dependent on the data mapping to the transformation space and image features. Entropy-based kernel graph cut method is suitable for segmentation of textured images. However, the quality of its segmentation is affected by the quality of extracting kernel centers. This paper examines the segmentation of textured images using the entropy-based kernel graph cut method based on weighted k-means. Using the advantages of kernel space, the objective function consists of two data terms to transfer the data standard deviation of each area in the segmented image and the regularization term. The proposed method, while using the advantages of suitable computational load of graph cut methods, will be a suitable alternative for segmenting textured images. Laboratory results have been taken on a set of well-known datasets that include textured shapes in order to evaluate the efficiency of the algorithm compared to other states-of-the-art methods in the field of kernel graph cut.