The Challenge

Since its invention in the early 1970’s, computed tomography (CT) has become one of the dominant approaches to non-invasive diagnostic imaging.  Yet despite its tremendous utility, there is a continuing desire to push the limits of detection, while minimizing the overall radiation exposure to the patient.  One approach to accomplishing this objective is to alter the image formation process.  One of the leading researchers in this field is Dr. Homer Pien at Massachusetts General Hospital. According to Dr. Pien, in contrast to conventional filtered back–projection image reconstruction algorithms, various research groups are looking into the potential for iterative reconstruction algorithms to substantially reduce radiation dosage while improving image signal-to-noise ratio and resolution.  The major impediment to iterative reconstructions, however, is the vast amount of computation necessary to form high quality images.  His research group (the Laboratory for Medical Imaging and Computations, Department of Radiology, Massachusetts General Hospital), utilizes a variety of computing platforms to prototype and validate their iterative reconstruction algorithms

The Solution

NZO Cloud delivered a High Performance Private Cloud Instance integrated with NVIDIA Telsa Graphical Processor Units (GPU).  The High Performance Private Cloud Instancewas delivered as a turn key solution with all necessary hardware and a custom developed software platform.  This allowed Dr. Pien to  simply install his own software packages and begin running jobs immediately.   

IMPACT

Their high performance computing solution from NZO Cloud provides a stable platform on which they can rapidly implement their algorithms using standard high-level programming languages (e.g., C/C++) and MPI standards. The NZO Cloud computing cluster architecture provides the inherent parallelism which permits their code to execute in a reasonable amount of time. 

ORGANIZATION PROFILE

The Laboratory for Medical Imaging and Computations is dedicated to improving the quality of clinical medical imaging systems through the use of sophisticated image processing and computational techniques. As part of the Massachusetts General Hospital’s Department of Radiology, LMIC serves as a laboratory where clinical imaging problems can find engineering solutions arising from the fields of signal/image processing, electrical engineering, computer science, and computer engineering.

NZO Cloud delivered a High Performance Private Cloud Instance integrated with NVIDIA Telsa Graphical Processor Units (GPU).

Staff Doctor Harvard Medical Imaging