High performance cluster computing in imaging research



Many advanced MRI techniques involve numerous computational complex processing steps that must be applied before the data can be passed to a biostatistician for final analysis. Performing these computations in a standardized and automated fashion is an integral part of unbiased and reproducible research.

We are dedicated to creating, managing, and running specialized processing pipeline software for large-scale imaging studies with high computational demands in preclinical and clinical settings. Our software is completely independent of the computation methods used and is able to handle multi-step processing in SLURM-based high-performance computation clusters, such as UB's Center for Computational Research HPC. Our software has been an essential part of several studies conducted by different research leaders, including the Buffalo Neuroimaging Analysis Center and the Center for Biomedical Imaging.

Publications


Der Medical Computation Server


Bw Lehr, Ferdinand Schweser, A. Deistung, D. Güllmar, J. Reichenbach


2011


MeCS – Integrating prototype processing programs into clinical routine


B. W. Lehr, F. Schweser, A. Deistung, D. Güllmar, J. Reichenbach


2010