GPU-based motion correction of contrast-enhanced liver MRI scans: An OpenCL implementation

Jihun Oh, Diego Martin, Oskar Skrinjar
School of Electrical and Computer Engineering, Georgia Tech, Atlanta, 30332, USA
IEEE International Symposium on Biomedical Imaging: From Nano to Macro, 2011


   title={GPU-based motion correction of contrast-enhanced liver MRI scans: An OpenCL implementation},

   author={Oh, J. and Martin, D. and Skrinjar, O.},

   booktitle={Biomedical Imaging: From Nano to Macro, 2011 IEEE International Symposium on},





Source Source   



Clinical diagnosis and quantification of liver disease have been improved through the development of techniques using contrast-enhanced liver MRI sequences. To qualitatively or quantitatively analyze such image sequences, one first needs to correct for rigid and non-rigid motion of the liver. For motion correction of the liver, we have employed bi-directional local correlation coefficient Demons, which is a variation of the original Demons method. However, despite the intrinsic speed of the Demons method, the run-time on the order of an hour of its CPU-based implementation is not sufficiently short for a regular clinical use. For this reason we implemented the method on a graphics processing unit (GPU) using OpenCL. On NVIDIA GTX 260M, which is a laptop GPU, we achieved sub-minute runtime for the motion correction of typical liver MRI scans, which was ~50 times faster than its CPU-based implementation. A sub-minute runtime of liver MRI motion correction allows for its regular clinical use.
No votes yet.
Please wait...

* * *

* * *

HGPU group © 2010-2017 hgpu.org

All rights belong to the respective authors

Contact us: