Related Experiment Video
Updated: Jun 8, 2026
09:41
A Pipeline for 3D Multimodality Image Integration and Computer-assisted Planning in Epilepsy Surgery
Published on: May 20, 2016
GPU-accelerated elastic 3D image registration for intra-surgical applications.
Daniel Ruijters1, Bart M ter Haar Romeny, Paul Suetens
1Interventional X-Ray Innovation, Philips Healthcare, The Netherlands. danny.ruijters@philips.com
Computer Methods and Programs in Biomedicine
|October 19, 2010
Summary
Accelerating elastic image registration for medical treatments using graphics processing units (GPUs) significantly reduces computation time. This GPU-accelerated method achieves a 50x speedup, enabling faster interventional procedures.
Area of Science:
- Medical Imaging
- Computational Biology
- Computer Science
Background:
- Elastic image registration is crucial for compensating local motion in intra-patient biomedical images.
- B-spline based elastic registration is computationally intensive, limiting its use in real-time interventional treatments.
Purpose of the Study:
- To accelerate B-spline based elastic image registration using graphics processing units (GPUs).
- To enable efficient calculation of similarity measures and their derivatives on the GPU.
Main Methods:
- Leveraged GPU parallel processing power for B-spline evaluation using hardware-accelerated tri-linear interpolation.
- Implemented a two-pass approach for calculating similarity measures and derivatives on the GPU.
Main Results:
- Achieved a significant speedup factor of approximately 50 compared to traditional CPU implementations.
- Demonstrated the feasibility of GPU acceleration for complex elastic registration tasks.
Conclusions:
- GPU acceleration dramatically reduces computation time for B-spline elastic image registration.
- This acceleration facilitates the application of elastic registration in time-critical interventional procedures.
