Research On Automatic Image Segmentation By Integrating Star Shape Prior

Extracting foreground objects of interest from the complex background is of great practical significance in the research of computer vision, pattern recognition and digital image processing. In general, the existing algorithms can be divided into two categories: interactive and automatic methods. Graph Cuts is an interactive method which is widely used in research and application. The main tasks of this thesis include the following parts:The first part of this paper focus on the research status and significance of image segmentation;The second part of the thesis studies the salient region detection technology. Mainly introduced several typical image salient region extraction method, mainly include the Itti algorithm, FT algorithm, HC algorithm and RC algorithm;The third part of the paper, we firstly introduce the Graph Cuts segmentation algorithm, next we introduce the star shape prior which is integrated with the Graph Cuts method;Finally, In order to extend the interactive image segmentation algorithm based on star shape constraints to an automatic approach. In order to achieve this goal, we design a strategy to automatically locate the star centers of the object that has star shape.A large numbers of comparative experiments on the publicly available data sets demonstrate the superior performance of our proposed method.