Development of Three Dimensional Bone Pattern Descriptors: Local Binary Patterns

摘要：

Three dimensional (3D) bone structures are described conventionally by a mode) of rod and plate descriptors. However, subtle bone growth or resorption process occurs in the early stage of bone diseases or dental implants and may not be detected by using this conventional model. In this paper, we describe a 3D local binary pattern (3dLBP) descriptor based on a two dimensional local binary pattern (LBP) descriptor that has been used quite successfully in machine vision applications such as texture classification and face recognition. The 3dLBP descriptor was applied to two synthetic images to understand the correspondence of individual 3dLBP features with geometric features. Then, the 3dLBP descriptor was applied to 3D human mandible images from cone beam computed tomograph (CBCT) to measure the difference between the cortical and trabecular bone patterns. 3dLBP descriptors detected more dark spots and less uniformity in cortical bones but more corners and edges in trabecular bones (Wilcoxon pairwise signed rank test, p<0.05). Since 3dLBP descriptor values were consistent with known geometric differences between cortical and trabecular bones, we conclude that 3dLBP descriptors should be useful in 3D bone pattern analysis.

关键词： three dimensional texture descriptor local binary pattern cone beam computed tomography trabecular bone microarchitecture bone remodeling

作者: Ashwini Kumar Mishra Desok Kim Ilma Andayana

作者单位: Department of Electronics and Electrical Engineering,IIT Guwahati,Guwahati, India Department of Electrical Engineering,KAIST,Daejeon, South Korea

会议类型: 国际会议

会议名称: 2011 4th International Congress on Image and Signal Processing(第四届图像与信号处理国际学术会议 CISP 2011)

会议地点: 上海

会议语种:英文

页码: 1905-1909

在线出版日期: 2011-10-15（万方平台首次上网日期，不代表论文的发表时间）