Continued Development and Maintenance of a Computational Morphometry Toolkit

计算形态测量工具包的持续开发和维护

• 批准号: 7626623
• 负责人: TORSTEN ROHLFING
• 金额: $ 12.77万
• 依托单位: SRI INTERNATIONAL
• 依托单位国家: 美国
• 财政年份: 2009
• 资助国家: 美国
• 起止时间: 2009-04-10 至 2011-03-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/7626623
• 关键词: AIDS/HIV problem; Aging; Alcohol abuse; Alcoholism; Algorithms; Alzheimer' s Disease; Angiography; Animals; Arts; Atlases; Biological; Brain; Breast; Cephalic; Chronic; Code; Communication; Communities; Complement; Complex; Computer software; Control Groups; Copyright; Data; Data Analyses; Data Set; Databases; Dementia; Detection; Development; Disease; Documentation; Drosophila genus; Ensure; Environment; Evolution; Future; Goals; Gold; High Performance Computing; Human; Image; Imaging Device; Insecta; Institution; Investigation; Libraries; Linear Models; Link; Maintenance; Malignant Neoplasms; Manuals; Maps; Modeling; Motion; Names; Neurosciences; Operating System; Operative Surgical Procedures; Partner in relationship; Patients; Performance; Procedures; Process; Publishing; Recipe; Reproducibility; Research Personnel; Schizophrenia; Science; Software Design; Software Tools; Source; Source Code; Speed; Statistical Models; Stroke; Structure; System; Techniques; Testing; Three-Dimensional Imaging; Time; Writing; age effect; base; bioimaging; clinically relevant; computational morphometry; computerized; design; flexibility; graphical user interface; image processing; image registration; imaging modality; improved; innovation; insight; interoperability; morphometry; open source; parallel computing; prevent; public health relevance; restoration; software systems; tool; usability; wiki

DESCRIPTION (provided by applicant): Over the past ten years, the PI has developed a flexible set of software tools for accurate computational morphometry, named the Computational Morphometry Toolkit (CMTK), which has been used in numerous studies, ranging from insect brain connection mapping to investigations of effects of alcoholism and aging on the human brain. CMTK's registration tools have also been successfully applied to extra-cranial problems, such as patient motion correction in contrast-enhanced MR breast imaging and CT subtraction angiography. The proposed project will make CMTK publicly available, improve its usability, and ensure its future support and development. The existing software consists of three parts: (1) a library with application-proven implementations of a large number of image processing and analysis algorithms; (2) command line tools for image restoration, registration, segmentation, and statistical modeling; (3) scripts for complex combinations of the elementary registration and segmentation tasks, such as longitudinal and cross-sectional morphometric studies. CMTK has a unique set of features, as it combines robust tools for image registration, segmentation, filtering, and modeling. These tools can be applied to data from different imaging modalities, and are useful for analysis of large, under-analyzed archival data sets. The software is end-user focused and computationally efficient, enabled by state-of-the-art parallel computation techniques. We propose three specific aims: (1) Access. Legacy third-party code in the existing software that prevents open source distribution will be replaced with freely available alternatives. (2) Usability. We will improve and document the pre-defined processing pipelines for complex studies (e.g., longitudinal deformation-based morphometry with a general linear model) and provide a user manual. (3) Maintenance. We will review and complete source code documentation, replace the current configuration and build environment, and set up auto- mated testing and bug tracking with state-of-the-art systems that are publicly available. We will improve interoperability with selected related soft- ware development efforts, which provide complementary functions to CMTK. (4) Enhancement. We will port key algorithms to state-of-the-art high-performance computing hardware. We will also implement an image-relationship database for transformation bookkeeping. By making publicly available a proven software system for morphometric studies, this project will enable any institution with appropriate image data to perform data analysis with efficiency, reliability, and reproducibility. Once released as the result of this project, CMTK will continue to enable new scientific insights in many fields of neuroscience, from human applications to insect science. The state-of-the-art registration tools in CMTK will furthermore continue to support clinically relevant applications such as treatment target delineation and patient motion correction. PUBLIC HEALTH RELEVANCE: Computerized morphometric analysis based on three-dimensional imaging enables high-throughput studies of disease (e.g., alcohol abuse, HIV/AIDS, stroke, dementia, and cancer), normal development, and aging effects on human and animal brain and body structure. Structural differences between diseased and control groups can be used to help identify potential biological causes of such conditions as Alzheimer's disease or schizophrenia. This project will enhance a proven soft- ware system for imaging-based morphometric studies and make it freely avail- able to the scientific community.

描述（由申请人提供）：在过去的十年中，PI开发了一套灵活的软件工具，用于精确的计算形态测量，称为计算形态测量工具包（CMTK），已用于许多研究，从昆虫大脑连接映射到酒精中毒和衰老对人脑影响的研究。CMTK的配准工具也已成功应用于颅外问题，例如对比增强MR乳腺成像和CT减影血管造影中的患者运动校正。拟议的项目将使CMTK公开可用，提高其可用性，并确保其未来的支持和发展。现有的软件由三部分组成：（1）一个库，具有大量图像处理和分析算法的应用程序验证的实现;（2）用于图像恢复、配准、分割和统计建模的命令行工具;（3）用于基本配准和分割任务的复杂组合的脚本，例如纵向和横截面形态测量研究。CMTK具有一组独特的功能，因为它结合了用于图像配准、分割、过滤和建模的强大工具。这些工具可应用于不同成像模式的数据，并可用于分析大型、分析不足的存档数据集。该软件以最终用户为中心，计算效率高，采用最先进的并行计算技术。我们提出了三个具体目标：（1）准入。现有软件中阻止开源分发的遗留第三方代码将被免费提供的替代品取代。(2)可用性。我们将改进和记录复杂研究的预定义处理管道（例如，具有一般线性模型的基于纵向变形的形态测量），并提供用户手册。(3)上维护我们将审查和完成源代码文档，替换当前的配置和构建环境，并使用公开可用的最先进的系统设置自动测试和错误跟踪。我们将通过选定的相关软件开发工作来提高互操作性，这些软件为CMTK提供了补充功能。(4)增强。我们将把关键算法移植到最先进的高性能计算硬件上。我们还将实现一个用于转换簿记的映像关系数据库。通过公开提供经过验证的形态学研究软件系统，该项目将使任何拥有适当图像数据的机构能够高效、可靠和可重复地进行数据分析。一旦作为该项目的成果发布，CMTK将继续在神经科学的许多领域实现新的科学见解，从人类应用到昆虫科学。此外，CMTK中最先进的配准工具将继续支持临床相关应用，例如治疗靶点描绘和患者运动校正。公共卫生相关性：基于三维成像的计算机化形态测定分析使疾病的高通量研究成为可能（例如，酒精滥用、艾滋病毒/艾滋病、中风、痴呆和癌症）、正常发育以及衰老对人类和动物大脑和身体结构的影响。患病组和对照组之间的结构差异可以用来帮助确定阿尔茨海默病或精神分裂症等疾病的潜在生物学原因。该项目将增强一个经过验证的基于图像的形态测量研究软件系统，并使其免费提供给科学界。