| With the development of computers, generally, all kinds of processing operations now require to satisfy the users with processing speed to improve user experience. As a basic and important subject in computer vision, image segmentation by now has been widely researched. From the former adaptive image segmentation to the present image segmentation based on user interactive, with the upgrading of segmentation algorithm, the result of image segmentation is becoming better and better. But many present image segmentation algorithm are highly complicated and they often ignore the processing speed of image segmentation on the bases of pursuing segmentation quality, resulting in not high user experience.The paper proposes the interactive image segmentation based on real-time feedback with its algorithm based on geodesic distance image segmentation. On the bases of geodesic distance segmentation, we by turn add in hierarchical image processing and local image processing method to the algorithm, and finally, use the algorithm of multiple threads to process image segmentation algorithm to make a real-time feedback allowing users to have a smooth experience.Its main contents and innovation points are:1. It first briefly introduces image segmentation, mainly explaining what is image segmentation, and its application situations, algorithm classification and development sate, and then briefly introduces the differences and connections between segmentation and matting.2. It gives a detailed introduction of interactive image segmentation, focusing on real-time segmentation based on tool of paint selection. Paint selection algorithm can also be called a tool whose basic working principle is very easy which is to successively process image segmentation with simple user interactive. It finally analyses the advantages and disadvantages of paint selection algorithm. The core of image segmentation algorithm based on paint selection still is the image segmentation approach based on Graph Gut. In this paper, our research is based on real-time feedback interactive image segmentation so that we use faster image segmentation algorithm based on geodesic distance image segmentation.3. Finally, we propose a real-time feedback interactive image segmentation which is mainly the image segmentation approaches based on paint selection and geodesic distance. We improve the image segmentation based on geodesic distance, accelerating the segmentation speed and at the same time improve the quality of image segmentation to an extent. |
