Bi-Textured Image Segmentation Based On Gabor Filter And Active Contour Algorithm

Texture is an important character of images, textured images exist extensively in the nature. Because of variety and complexity of texture itself, there are a lot of methods to extract the character of the texture. All the while, the analysis and process of textured images is research hotspot of image process technology.Textured image segmentation is of great significance not only to many researches of computer vision and image process but also to practical application. Textured image segmentation comes down to all kinds of images and appears in all fields of image process. There are a lot of applications such as cancer cell recognition in medical images and object identification in remote sense images.Textured image segmentation is to divide an textured image to different regions, such that there is only a kind of texture. Base on different methods of extract character of texture, there are three methods: statistics method, structure method and spatial/frequency method. Spatial/frequency method is a new method developed in the last decade. Its most typical form is to adopt Gabor filter to extract character of image, then textured image segmentation is changed to general image segmentation. Active contour algorithm presented by Kass is a method studied and applied extensively in image segmentation. This method is superior to traditional method of image segmentation.This paper combines the method of optimal Gabor filter design presented by Wu GaoHong and Zhang YuJin and multichannel active contour model presented by Chan. Based on these, a new method which is based on Gabor filter and multichannel active contour algorithm is proposed. Experiment proves that new method made good use of character of both texture and covered the shortage of former method.