Parallel Gaborfilter Based Texture Analysis

Texture is universal and Complex diversity,there is not a satisfactory method worked out in texture analysis field,and it engages people research further more in this field.Up till now, Image texture analysis is a hot topic in field of study at home and abroad.This paper selects Gaborfliter in image texture analysis.Gaborfilter is so good at texture analysis that making it very suitable for texture analysis.Certain questions about Real-time Performance and Adaptability exsit when processing images.This article uses Parallel Computing to make further studies. It helps to short the time of image processing in Real-time Performance, and in Adaptability,this paper proposes a new method in Parameter Settings,it helps to short the time of image processing , promote Gabor filtering efficiency and widen the application of Gaborfilter. There are some main conclusions of this paper:1) Do some research on Parallel Computing and Image texture analysis. Meanwhile, introduce and study the importance and the analytical method of texture analysis,compare existing methods of Image texture analysis.2) Improve gabor filter by using Parallel Computing,and make the paralleled Gaborfilter faster in image filtering by GPU.3) Proposes a new method in Parameter Settings and raise Gaborfilter's applicability.4) Test the improved method and analyze the effect of Parallel Computing. At the same time, analyze and apply the image that has been processed by Gabor filter in order to recognize images.