A Novel Texture Descriptor: Circular Parts Local Binary Pattern

Local Binary Pattern (LBP) are considered as a classical descriptor for texture analysis, it has mostly been used in pattern recognition and computer vision applications. However, the LBP gets information from restricted number of local neighbors which is not enough to describe information, other descriptors that get large suffer dimensionality consume much time. In this regard, we propose novel classification known Circular Parts (CPLBP) designed enhance by extending area neighborhood one region using polar coordinates permit capture more discriminating relationships exists amongst pixels increase efficiency extracting features. Firstly, circle divided into regions with specific radius angle. After that, calculate average gray-level value each part. Finally, center pixel compared these values. The relevance proposed idea validate databases Outex 10 12. A complete evaluation on benchmark data sets reveals CPLBP's high performance. CPLBP generates score 99.95 SVM classification.