Interactive Image Segmentation Based On Graph Cuts

Abstract:With the development of technology of computer vision and digital image processing, image segmentation has becomes vital step of image processing and image analysis. Because interactive segmentation can achieve more accuracy result comparing with automatic method, it becomes more and more popular. And at the same time, Graph Cuts with its excellent behavior in segmentation field, has receives more and more attraction.Recently, Super Pixel has been used in image preprocessing, for that it can capture redundancy in image and greatly reduce the complexity of subsequent image processing tasks, and also for its ability to preserve boundary to make segment results more like an object. So we form a Graph Cuts base on Super Pixel instead of pixel level, which makes a great reduce in the number of graph nodes and make the Graph Cuts more efficient. And we present a novel interactive image segmentation method based on the initial segmentation of improved watershed.In our algorithm, firstly, the interactive information is introduced as markers, which are input by the users to roughly indicate the position and main features of the object and background. The markers can be the simple strokes, and they should not be overlapped. Then the modified watershed was used to initialize segmentation. Second, the small regions, instead of pixels, are seen as nodes to construct the graph. The cost function is composed of the region properties and the boundary properties, and the energy minimization can be done with the maximum-flow/minimum cut. Finally, the boundary regions are refined by incorporating local information of pixels and global information of regions.