Snake Model Based On Minimum Spanning Tree Algorithm For Image Segmentation To Improve

Image segmentation is based on some characteristics of the image, has certain properties similarity in accordance with together, to split the difference between different partitions to large areas of pixels. We already have a variety of classic segmentation algorithms, but still not a segmentation method can generate all the images are ideal for the segmentation results. Based on the different kind of method of image segmentation, according to the characteristics of the feature, for segmentation images with specific image characteristics of research, become improve image segmentation method an important means of image segmentation effect.Image segmentation based on graph theory is developed in recent years, a new image segmentation method to get people's attention. This paper introduces the basic concept of graph theory, gives the image the basic method of mapping a graph, based on visual characteristics define the similarity between adjacent pixels. Minimum spanning tree based on the theory of image segmentation by snake model has been optimized to overcome the traditional snake model, the sensitivity of the initial position, improve the ability to extract the image edge sag. Finally given based on minimum spanning tree snake model and the image segmentation algorithm are proposed. Experiments show that this algorithm has better image segmentation accuracy.