CAREER: Parallel Feature Extraction and Dataset Registration for Volumetric Visualization

体图像中结构的提取是可视化中的重要早期任务。 在结构提取之后，结构可以被渲染以供科学家或临床医生最终显示。 数据集登记对于监测变化和评估干预策略也很重要。 这个项目将侧重于层析数据的可视化，虽然开发的方法可能是有用的其他体积数据域的相关问题。该项目将在其他（非医学）体积数据集上对PI的方法进行经验测试，以检查方法的一般实用性。 在传统的医学应用中，断层扫描数据尚未得到充分利用。 这项研究的一个关键部分是支持更有效地利用体积数据。 特别是，对于医疗应用，本提案中提出的可视化阶段必须近实时完成，以支持诊断、治疗计划、干预指导和治疗评估。

Extraction of structures in volumetric images is an important early task in visualization. Following structure extraction, the structures can be rendered for final display for scientists or clinicians. Dataset registration is also important for monitoring change and assessing intervention strategies. This project will focus on visualization of tomographic data, although the methods developed are likely to be useful for related problems in other volumetric data domains. The project will empirically test the PI's methods on other (non medical) volumetric datasets to examine the general utility of the methods. In traditional medical applications, tomographic data has not been thoroughly exploited. A key part of this research is support of more effective utilization of volumetric data. In particular, for medical application, the visualization stages addressed in this proposal must be accomplished in near real-time to support diagnosis, treatment planning, intervention guidance, and treatment assessment.