Home
Scholarly Works
GPU implementation of a deformable 3D image...
Journal article

GPU implementation of a deformable 3D image registration algorithm

Abstract

We present a parallel implementation of a new deformable image registration algorithm using the Computer Unified Device Architecture (CUDA). The algorithm co-registers preoperative and intraoperative 3-dimensional magnetic resonance (MR) images of a deforming organ. It employs a linear elastic dynamic finite-element model of the deformation and distance measures such as mutual information and sum of squared differences to align volumetric image data sets. Computationally intensive elements of the method such as interpolation, displacement and force calculation are significantly accelerated using a Graphics Processing Unit (GPU). The result of experiments carried out with a realistic breast phantom tissue shows a 37 fold speedup for the GPU-based implementation compared with an optimized CPU-based implementation in high resolution MR image registration. The GPU implementation is capable of registering 512 × 512 × 136 image sets in just over 2 seconds, making it suitable for clinical applications requiring fast and accurate processing of medical images.

Authors

Mousazadeh H; Marami B; Sirouspour S; Patriciu A

Journal

Conference proceedings: ... Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual Conference, Vol. 2011, , pp. 4897–4900

Publication Date

December 26, 2011

DOI

10.1109/IEMBS.2011.6091213

ISSN

1557-170X
View published work (Non-McMaster Users)

Contact the Experts team