RESOURCE FOR BIOMEDICAL COMPUTING

本修订提案要求为BRTP生物医学项目提供4年支持 强调定量生物医学成像的计算资源。 目标是1）开发对健康重要的新计算技术- 相关研究，2）与生物医学科学家合作解决 需要高级计算技术的问题，3）使资源 研究人员可获得的能力，4）培训科学家， 生物医学计算，以及5）开发和领域有用的生物医学- 计算系统，本地和全国。 技术研发跨越理论发展到实施 系统的输出。 理论发展包括建模 生物现象和数据采集的转换。 的 计算方法的发展包括1）建模加 从测量中提取图像信息的算法开发 数据，2）高性能处理器的应用，以及3） 开发可快速运输的系统， 整合到我们的合作者和更广泛的 科学界。 核心研究的具体领域包括：1） 通过改进正电子发射断层摄影的算法 衰减校正，将其他来源的信息应用于图像 估计，并将当前方法扩展到三维数据，2） 建模和图像估计方法，用于改善扫描和 计算光学切片显微镜，和3）新的方法， 全局形状建模和图像表示，以适应多个 多维生物医学图像中的可变性水平。 技术发展包括1）SIMD和MIMD的开发 用于获得有用性能的并行处理方法 计算上具有挑战性的算法，以及2）高速通信 用于将资源能力分布到分散的站点的系统， 技术和生物医学研究。 一个集成环境 用于定量成像研究。 合作研究包括9个校内（华盛顿大学）和 9个校外项目：功能性脑映射，认知 神经科学，心肌和肿瘤成像，图像统计方法 分析，通过MRI的脑血流，多细胞形态发生， 胚胎肌动蛋白结合蛋白，突触后受体可塑性，高 分辨率荧光显微镜，抗HIV寡核苷酸，模式- 理论方法，变形拓扑模板，放射治疗 计划，眼底成像，基于ATM的网络，遗传连锁 分析，心肌缺血的细胞生物化学，散射校正 用于PET、DNA绘图和电子显微镜放射自显影分析。 四种资源技术为10个校内和5个 校外研究项目

This revised proposal requests 4-year support for a BRTP Biomedical Computing Resource that emphasizes quantitative biomedical imaging. Goals are to 1) develop new computing technologies important to health- related research, 2) collaborate with biomedical scientists to solve problems requiring advanced computing techniques, 3) make Resource capabilities accessible to investigators, 4) train scientists in biomedical computing, and 5) develop and field useful biomedical- computing systems, locally and nationally. Technological R&D spans theoretical developments through implementation and exportation of systems. Theoretical developments include modeling of biological phenomena and of transduction for data acquisition. The development of computing methodologies includes 1) modeling plus algorithm development for abstracting image information from measurement data, 2) the application of high-performance processors, and 3) the development of systems that can be expeditiously transported and integrated into the environments of our collaborators and the broader scientific community. Specific areas of core research include 1) algorithms for advancing positron-emission tomography by improving attenuation corrections, applying information from other sources to image estimations, and extending current methods to 3-dimensional data, 2) modeling and image-estimation methods for improving both scanning and computational optical-sectioning microscopy, and 3) new approaches to global shape modeling and image representation to accommodate multiple levels of variability in multi-dimensional biomedical images. Technological development includes 1) exploitation of both SIMD and MIMD parallel-processing methods for achieving useful performance of computationally challenging algorithms, and 2) high-speed communication systems for distributing Resource capabilities to dispersed sites for both technological and biomedical research. An integrated environment for quantitative-imaging research is maintained. Collaborative research includes 9 intramural (Washington University) and 9 extramural projects in: functional brain mapping, cognitive neuroscience, myocardial and tumor imaging, statistical methods in image analysis, cerebral blood flow via MRI, multicellular morphogenesis, embryonic actin-binding proteins, postsynaptic-receptor plasticity, high- resolution fluorescence microscopy, anti-HIV oligonucleotides, pattern- theoretic methods, deformable topologic templates, radiation-treatment planning, ocular-fundus imaging, ATM-based networking, genetic-linkage analysis, cellular biochemistry in cardiac ischemia, scatter corrections for PET, DNA mapping, and electron-microscopic autoradiograph analysis. Four Resource technologies provide user service for 10 intramural and 5 extramural research projects.