Research On Theory And Application Of Local Binary Patterns

Texture analysis is a process of extracting texture parameters by a certain imageprocessing techniques to obtain qualitative description of textures.In recent years,the textureanalysis method based on local binary pattern （LBP） is considered as one of thestate-of-the-art methods for its simple structure and superior performance.Instead of trying toanalysis texture on a pixel level,local binary pattern uses local patterns to achieve it.Thispaper makes a more in-depth study on this texture analysis method.According to the theory ofLBP and its wide range of applications, this paper will be divided into three parts:First,by reading a lot of literature and in-depth study on LBP,we make a systematicanalysis and summary of it.We make a profound analysis of the classic local binary pattern asan example,and make a induction of its’ shortage.Then we make a review.of the application oflocal binary patterns in fields of image processing.Second,we make a in-depth study on the theory of steerable filters and design a filter withsuperior performance.This paper uses the second derivative of Gaussian function to constructa superior performance of steerable filters,and use it to accomplish the work of imagesegmentation.At last,a new texture extraction method was proposed with the steerable filtersand local binary patterns which called local steerable binary patternsmodel(LSBP).Experiments show that the new model achieve good results.Finally,we study the work of Face Recognition based on LSBP.We use LSBP to extractface image information to realize the work of face recognition.The experimental results showthis method achieve a better recognition performance.