A Texture Extraction Method Based on Local Binary Pattern Operator
Xing San, Xinmei Tian, Xiuqing Wu · 2006
Most of image texture extraction methods are of high computational complexity, which hardly restricts their application in image processing fields. This paper proposes an improved local binary pattern operator to extract the image texture features. In the algorithm, single pixel is replaced with homogeneous object obtained by an object-oriented image segmentation method to exact image texture. At the same time, conditional probability is adopted as the parameter of image texture. The improvement makes it more valid to analysis object by fuzzy inference based image textures. In our experiments, the extracted image texture features are utilized for classifying images. The achieved good results indicate that the proposed method is faster than other methods while remaining close classification performance.