Joint-scale LBP: a new feature descriptor for texture classification

详细信息    查看全文

• 作者：Xiaosheng Wu ; Junding Sun
• 关键词：Texture classification ; Local binary pattern (LBP) ; Joint ; scale local binary pattern (JLBP) ; Complete JLBP (CJLBP)
• 刊名：The Visual Computer
• 出版年：2017
• 出版时间：March 2017
• 年：2017
• 卷：33
• 期：3
• 页码：317-329
• 全文大小：
• 刊物类别：Computer Science
• 刊物主题：Computer Graphics; Computer Science, general; Artificial Intelligence (incl. Robotics); Image Processing and Computer Vision;
• 出版者：Springer Berlin Heidelberg
• ISSN：1432-2315
• 卷排序：33

文摘

This paper presents a simple, efficient, yet robust approach, named joint-scale local binary pattern (JLBP), for texture classification. In the proposed approach, the joint-scale strategy is developed firstly, and the neighborhoods of different scales are fused together by a simple arithmetic operation. And then, the descriptor is extracted from the mutual integration of the local patches based on the conventional local binary pattern (LBP). The proposed scheme can not only describe the micro-textures of a local structure, but also the macro-textures of a larger area because of the joint of multiple scales. Further, motivated by the completed local binary pattern (CLBP) scheme, the completed JLBP (CJLBP) is presented to enhance its power. The proposed descriptor is evaluated in relation to other recent LBP-based patterns and non-LBP methods on popular benchmark texture databases, Outex, CURet and UIUC. Generally, the experimental results show that the new method performs better than the state-of-the-art techniques.