The proliferation and increased reliance on high-resolution, multimodality biomedical images/videos present significant challenges for a broad spectrum of clinical practitioners and investigators throughout the clinical and research communities. A growing spectrum of new and improved imaging modalities and new image processing techniques can significantly affect diagnostic and prognostic accuracy and facilitates progress in the areas of biomedical research and discovery. However, the impact of these new technologies in both time-critical clinical applications and high-throughput research pursuits depends, in large part, on the speed and reliability with which the imaging data can be visualized, analyzed and interpreted. Conventional serial computation is grossly inadequate and inefficient for managing; these increasing amounts of data and the employment of the new advances in medical imaging is often limited by insufficient compute and storage resources.
High performance computing (HPC) ranging from multi-core CPUs and GPU-based processing to parallel machines to Grids and Clouds are effective mechanisms for overcoming such limitations. They allow for significant reduction of computational time for running large experiments and speed-up the development time for new algorithms while increasing the availability of new methods for the research community, and supporting large-scale multi-centric collaborations.
The workshop will build on existing collaborative efforts in understanding current trends in HPC medical imaging research. It will demonstrate and encourage open discussion regarding the current status and latest developments in the field; explore new ideas/motifs, identify the challenges which currently impeded wider adoption of these technologies in image-assisted translational research, clinical intervention and decision-making, and present innovative solutions to the challenges.
The intersection of medical imaging and high performance computing has demonstrated the potential to make a significant impact in the translation of medical imaging research from the lab into clinical practice. However, high-performance computing technologies are complex and their exploration in the medical domain is leading to new and interesting research challenges. Within the last years, a new generation of heterogeneous computing platforms consisting of nodes with multi-core CPUs and GPUs and deep memory hierarchies have emerged and is now being widely deployed for large scale scientific research. New trends like cloud infrastructures and interoperability between different HPCs augment the options to employ such infrastructures for medical imaging applications. Significant progress in security, reliability, availability and robustness of distributed IT infrastructures pave the way for their utilization not only in basic and clinical research but also in clinical routine. Nevertheless, many challenges in synthesizing and managing information from large quantities of high resolution imaging data remain. These challenges require novel techniques for data management, data retrieval and staging, and data processing and visualization to minimize I/O and computation overheads, exploit computing and memory capacity distributed across multiple types of devices and systems, and maximize application throughput while reducing application development complexity.
The academic goal of this workshop is to provide a high quality forum for researchers from different but directly related communities including academic, healthcare and industrial organizations to discuss and disseminate their latest findings, ideas and research results on these challenges.