Computational Modeling of Organ Structures

近年来，出现了一个名为“计算解剖学”的研究领域，它研究包含学科间变异的人体解剖学的数学模型。在这一优先领域，人体解剖学的数学模型被命名为“Dentai Shinsho”，并被视为实现智能辅助诊断的必要代表。在这个小组中，我们侧重于统计地图集，这是数学模型的一种典型形式，是基于对人口数据的统计分析而构建的。我们研究了统计地图集的构建框架及其应用。我们开发了(1)统计图谱的层次组织方法及其在从3D图像中自动分割和从2D图像中恢复3D形状中的应用；(2)针对层次图谱的构建，提出了精确重建器官内部结构的方法，例如分支管状结构的提取和基于分支管结构的区域分解；(3)相关的研究课题，包括统计图谱的基本工具和新的应用。

Recently, the research field called "Computation Anatomy" has emerged, in which mathematical models of human anatomy including inter-subject variations are investigated. In this Priority Area, the mathematical model of human anatomy was named "Dentai Shinsho" and regarded as an essential representation to realize intelligent assistance in diagnosis. In this group, we focused on statistical atlases, which is a typical form of the mathematical models and constructed based on statistical analysis of population data. We studied the framework of the construction of statistical atlases and their applications. We have developed (1) methods for hierarchical organization of statistical atlas and their applications to to automated segmentation from 3D images and 3D shape recovery from 2D images, (2) methods for accurate reconstruction of organ internal structures such as extraction of branching tubular structures and territory decomposition based on them aiming at construction of hierarchical atlases, and (3) related research topics including basic tools and novel applications of statistical atlases.