Resource Development for the Java Image Science Toolkit

Java 图像科学工具包的资源开发

• 批准号: 8013701
• 负责人: Bennett A. Landman
• 金额: $ 14.7万
• 依托单位: VANDERBILT UNIVERSITY
• 依托单位国家: 美国
• 财政年份: 2010
• 资助国家: 美国
• 起止时间: 2010-09-30 至 2012-09-29
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/8013701
• 关键词: Resource; Development; Java; Image; Science

DESCRIPTION (provided by applicant): Medical imaging data can be gathered in such vast quantities that it is often a challenge to carry out appropriate processing steps despite publicly available tools that might already exist for carrying out many of the procedures. A tool that automatically handles formatting and compatibility issues while integrating software from a variety of platforms is essential. The neuroimaging community benefits from several, excellent pipeline tools and image processing libraries. These environments provide end-users with systems to process and visualize large-scale datasets. While there has been substantial convergence in terms of the desirable features in a successful system (e.g., grid processing, block diagrams), the Java Image Science Toolkit (JIST) uniquely address important challenges. Cross-platform compilation and deployment are notoriously difficult, but JIST is based on the Java programming language (Sun Microsystems, Santa Clara, CA), which is inherently cross-platform and able to run on almost any platform. To date, no other pipeline software has made native use of neuroimaging image processing libraries with Java. JIST enables a unique flavor of "write-once, run many" development: programmers need only write a core section of code and users can access this functionality in many different ways (e.g., within block diagram, as a plugin, from the command line, from within Matlab, on a grid, etc.). Thus, JIST provide a seamless development path from prototype to cluster-based parallel processing. The proposed research and development effort will significantly improve the interoperability and adoptability of the JIST framework to enhance adoption by the broader neuroimaging research community. Specifically, this work will (1) enhance developers' ability to monitor and validate modular routines in collaborative development and (2) improve the usability through interactive visualization capabilities and detailed documentation. These aims have been chosen to direct development efforts at the specific concerns of existing JIST users. The primary hypothesis of this proposal is that by addressing the specific concerns of current JIST users, the platform will be made more accessible to the broader neuroimaging community. In turn, JIST will provide a more significant contribution toward advances in clinical research. These developments will ease the learning curve and provide more intuitive and responsive experiences for both developers and users. Users will be able to more readily leverage the substantial image analysis capabilities already available within JIST and benefit from improved accessibility of advanced features. PUBLIC HEALTH RELEVANCE: The proposed research improves the interoperability and adoptability of the Java Image Science Toolkit (JIST) resource using the Neuroimaging Informatics Tools and Resources Clearinghouse infrastructure. This effort will provide a more intuitive and responsive experience for both developers and users through an improved user interface and an automated algorithm validation and testing system. The end result will be that clinical investigators and image scientists will be able to more readily leverage the substantial analysis capabilities already available within JIST.

描述(由申请人提供)：医学成像数据可以如此大量地收集，以至于尽管可能已经存在用于执行许多程序的公开可用的工具，但执行适当的处理步骤通常是一个挑战。在集成来自各种平台的软件的同时，自动处理格式和兼容性问题的工具是必不可少的。神经成像社区受益于几个优秀的管道工具和图像处理库。这些环境为终端用户提供了处理和可视化大规模数据集的系统。虽然在成功的系统中的理想功能(例如，网格处理、框图)方面有很大的趋同，但Java图像科学工具包(JIST)是唯一解决重要挑战的工具。跨平台编译和部署是出了名的困难，但JIST是基于Java编程语言(Sun Microsystems，Santa Clara，CA)，该语言本质上是跨平台的，并且能够在几乎任何平台上运行。到目前为止，还没有其他管道软件在Java中原生使用神经成像图像处理库。JIST支持一次编写、多次运行的独特风格的开发：程序员只需编写代码的核心部分，用户可以通过许多不同的方式访问该功能(例如，在框图中、作为插件、从命令行、从MatLab内、在网格上等等)。因此，JIST提供了从原型到基于集群的并行处理的无缝开发路径。拟议的研究和开发工作将显著提高JIST框架的互操作性和可采纳性，以促进更广泛的神经成像研究社区的采用。具体地说，这项工作将(1)增强开发人员在协作开发中监控和验证模块化例程的能力，以及(2)通过交互式可视化功能和详细的文档来提高可用性。选择这些目标是为了指导针对现有综合技术用户的具体关切的开发工作。这项提议的主要假设是，通过解决当前JIST用户的具体关切，该平台将更容易为更广泛的神经成像社区所使用。反过来，JIST将为临床研究的进步做出更重大的贡献。这些开发将缓解学习曲线，并为开发人员和用户提供更直观、更具响应性的体验。用户将能够更容易地利用JIST内已有的大量图像分析能力，并受益于高级特征的更好的可及性。 公共卫生相关性：拟议的研究利用神经成像信息学工具和资源信息交换中心的基础设施，提高了Java图像科学工具包(JIST)资源的互操作性和可采用率。这一努力将通过改进的用户界面和自动化的算法验证和测试系统，为开发人员和用户提供更直观和反应迅速的体验。最终结果将是临床研究人员和影像科学家将能够更容易地利用JIST已有的实质性分析能力。

项目成果

Automatic Segmentation of Abdominal Wall in Ventral Hernia CT: A Pilot Study.

腹疝 CT 腹壁自动分割：一项试点研究。

• DOI: 10.1117/12.2007060
• 发表时间: 2013
• 期刊: Proceedings of SPIE--the International Society for Optical Engineering
• 作者: Xu,Zhoubing;Allen,WadeM;Poulose,BenjaminK;Landman,BennettA
• 通讯作者: Landman,BennettA

Quantitative Anatomical Labeling of the Anterior Abdominal Wall.

腹前壁的定量解剖标记。

• DOI: 10.1117/12.2007071
• 发表时间: 2013
• 期刊: Proceedings of SPIE--the International Society for Optical Engineering
• 作者: Allen,WadeM;Xu,Zhoubing;Asman,AndrewJ;Poulose,BenjaminK;Landman,BennettA
• 通讯作者: Landman,BennettA

Robust Optic Nerve Segmentation on Clinically Acquired CT.

• DOI: 10.1117/12.2043715
• 发表时间: 2014-03-21
• 期刊: Proceedings of SPIE--the International Society for Optical Engineering
• 作者: Panda S;Asman AJ;Delisi MP;Mawn LA;Galloway RL;Landman BA
• 通讯作者: Landman BA

Texture analysis improves level set segmentation of the anterior abdominal wall.

• DOI: 10.1118/1.4828791
• 发表时间: 2013-12
• 期刊: Medical physics
• 影响因子: 3.8
• 作者: Zhoubing Xu;W. M. Allen;R. Baucom;B. Poulose;B. Landman