This subproject is one of many research subprojects utilizing the resources provided by a Center grant funded by NIH/NCRR. The subproject and investigator (PI) may have received primary funding from another NIH source, and thus could be represented in other CRISP entries. The institution listed is for the Center, which is not necessarily the institution for the investigator. The Insight Toolkit (ITK) has become the de facto standard platform for advanced segmentation and registration research at many laboratories. At the same time, there is an increasing trend to deploy grid-computing infrastructures to support computations on extremely large data sets like those associated with the Visible Human Project. The architecture of ITK is not designed to support such efforts. We believe it is important to revisit and refine critical aspects of the architecture of ITK to support the emerging standards in the grid-computing community and to develop example applications to demonstrate the power of the ITK/grid combination in real-world research computing scenarios.
The specific aims as proposed in the application are: 1. Needs analysis Use task-level parallelization of large population data sets. 2. CPU architecture, instruction set, and cache size optimizations Examine optimizations of algorithms and inner loops for thread parallelism. 3. Message-passing parallelism in medical imaging algorithms in ITK Distribute individual jobs across nodes. 4. Grid protocols (MPI-GM) for distributed computations on the grid infrastructure Ongoing support of task, threading, and message-passing layers. We have identified two driving applications to provide real-world examples of the biomedical computing demands and constraints for this architecture engineering effort: namely, imageguided neurosurgery and image-guided radiofrequency ablation (RFA). Both applications rely heavily on computations that occur during interventional procedures, with hard realtime constraints. They also require extensive preoperative planning including image segmentation, registration, and sophisticated data analysis, each of which is amenable to task level parallelism executed on a grid platform. Currently, we are designing, implementing, and evaluating modifications of the ITK architecture and platform to better leverage grid computing in these scenarios.
Showing the most recent 10 out of 507 publications
