The Research On Texture Segmentation Of Ill Endoscopic Image

The properties of texture determine that the specific texture need different texture segmentation method.Therefore, In this paper, the purpose is to find an effective method to segment endoscopic image. The main research work are as follows:(1)Based on wavelet transform texture segmentable research,the paper uses wavelet transform to extract texture features. The k-means clustering algorithm is used to distinguish the texture characteristics.Finally, segmentation results are obtained.(2)The improvement of Gabor filter is that the number of the direction of Gabor filter is expanded from four to six(0°,30°,60°,180°,300°,330°).(3)Based on Gabor filter texture segmentable research,the paper describes the detail of design process of Gabor filter. After using the designed Gabor filter to extracte texture feature of endoscopic image,the paper uses k-means clustering algorithm to extract features.