Success Stories

Image Processing of MRI Brain Scans

Results: 9X Gain: 5 minutes vs 45 minutes

The Challenge

brain scan imagery

Magnetic resonance imaging (MRI) is widely used by researchers in a broad set of applications, such as distinguishing pathological tissue from normal tissue, studying brain function, and understanding correlations between brain structure and diseases.

MATLAB® is a popular desktop tool in neuroimaging for both algorithm development and production analysis. However, over the last several years, image processing has become increasingly computationally demanding due to the growing data volumes and the complexity of the processing algorithms. As new generations of MRI machines enable dramatic increases in resolution, speed of acquisition, volumetric imaging, and temporal studies, computing requirements of many neuroimaging applications have outstripped desktop workstations. More and more, scientists turn to HPC servers to support image processing workloads. However, porting applications to a parallel HPC environment is a specialized activity, and one which many neuroscience environments do not have access to.

Star-P Solution

With Star-P, neuroscientists can easily tap the power of parallel high performance computers to solve computational problems in a fraction of the time. Star-P lets researchers write applications on desktop computers using MATLAB®, and run them interactively on multi-processor servers. This eliminates the need to re-program the applications in C, FORTRAN or MPI languages in order to run on the parallel computer. Star-P's data parallel mode can be applied to processing of large images, and the task parallel mode can be applied to independent processing of multiple moderately-sized images.

Summary & Metrics

  • High-speed image processing application
  • FFT-based algorithms
  • No need to re-write MATLAB® code in C, Fortran, MPI
  • Perfusion analysis of 256x256 image that took 45 minutes on a desktop runs in under 5 minutes on a 4-processor SGI Altix server