Interactive Segmentation Algorithm Based on Improved Random Walk

Abstract

Interactive segmentation algorithm has been paid more and more attention in the field of image segmentation. Random walk algorithm is a proposed interactive segmentation algorithm based on graph theory, this paper proposes an improved Random Walk algorithm, the algorithm is more real-time than the original algorithm, and also has good robustness. This method adds the new improved Watershed algorithm to pre segmentation processing, reconstruct the connected domain graph, and the segmentation function Gauss area using 2 norm entropy reduction as a function of energy, finally it was solved by Random Walk algorithm. This method is faster than the original algorithm, the accuracy is satisfactory, and can be used for multi-target segmentation which has good robustness to noise.