@phdthesis{oai:kyutech.repo.nii.ac.jp:00004071,
 author = {Hossain, Shahera},
 month = {2014-07-30},
 note = {chapter 1 introduction||chapter 2 features for texture analysis||chapter 3 in-depth analysis of texture databases||chapter 4 analysis of features based on co-occurrence image matrix||chapter 5 categorization of features based on co-occurrence image matrix||chapter 6 texture recognition based on diagonal-crisscross local binary pattern||chapter 7 conclusions and future work, In this thesis, we focus on several important fields on real-world image texture analysis and recognition. We survey various important features that are suitable for texture analysis. Apart from the issue of variety of features, different types of texture datasets are also discussed in-depth. There is no thorough work covering the important databases and analyzing them in various viewpoints. We persuasively categorize texture databases ― based on many references. In this survey, we put a categorization to split these texture datasets into few basic groups and later put related datasets. Next, we exhaustively analyze eleven second-order statistical features or cues based on co-occurrence matrices to understand image texture surface. These features are exploited to analyze properties of image texture. The features are also categorized based on their angular orientations and their applicability. Finally, we propose a method called diagonal-crisscross local binary pattern (DCLBP) for texture recognition. We also propose two other extensions of the local binary pattern. Compare to the local binary pattern and few other extensions, we achieve that our proposed method performs satisfactorily well in two very challenging benchmark datasets, called the KTH-TIPS (Textures under varying Illumination, Pose and Scale) database, and the USC-SIPI (University of Southern California ― Signal and Image Processing Institute) Rotations Texture dataset., 九州工業大学博士学位論文 学位記番号:工博甲第354号　学位授与年月日:平成25年9月27日, 平成25年度},
 school = {九州工業大学},
 title = {Study on Co-occurrence-based Image Feature Analysis and Texture Recognition Employing Diagonal-Crisscross Local Binary Pattern},
 year = {}
}