期刊名称：Signal & Image Processing : An International Journal (SIPIJ)
印刷版ISSN：2229-3922
电子版ISSN：0976-710X
出版年度：2013
卷号：4
期号：1
页码：65
出版社：Academy & Industry Research Collaboration Center (AIRCC)
摘要：The method is based on recognizing that certain local binary patterns, termed are fundamental propertiesof local image texture and their occurrence histogram is proven to be a very powerful texture feature. TheLocal Binary Pattern (LBP) is a texture descriptor based on the probability of occurrence of elementarybinary patterns (texels) defined over a circular window. A new feature set derived from the LBP, called theLBP-Constant-Symmetry (LBP-CS) and LBP-High-Symmetry (LBP-HS) are proposed for recognition ofstone textures. The features are computed from each band of an isotropic color LBP Matrix for recognition.The tests were conducted in a variety of industrial samples. The obtained results are promising and showthe possibility of efficiently recognizing complex industrial products based on color and texture features
