Diagnostic imaging contributes substantially to the rapid proliferation of data in health-care centers. The increase has been attributed to two main factors:
- the aging population and concomitant increase in disease incidence and the required care, and
- the recent improvements in the technology of imaging which result in higher quality and therefore more voluminous data sets (A single imaging center can produce terabytes of images each year).
The proliferation of diagnostic images puts extra pressure on radiologists and affects their efficiency and reliability in reading images and handling cases. New methodologies and systems are required to assist the radiologists in coping with this trend and improving their performance.
Towards the goal of improving a radiologist's ability to accurately and efficiently cope with the proliferation of data, we are focusing on developing high-performance analytics including rigid and non-rigid registration and 3D reconstruction and visualization.
We have demonstrated a version of rigid registration running on the IBM Cell Broadband Engine which is approximately 49 times faster than an x86 implementation of the algorithm.
Our high-performance analytics are running on the Medical Analytics Platform which combines hardware, middleware, storage, and networking in a balanced configuration, thereby providing an efficient, scalable environment for managing, analyzing and displaying patient data.
Partnerships and Projects
- Mayo-IBM Medical Imaging Research
- Movie of the AALIMS System developed at the IBM Almaden Research Center.