TY - GEN
T1 - GPU-based stochastic-gradient optimization for non-rigid medical image registration in time-critical applications
AU - Bhosale, Parag
AU - Staring, Marius
AU - Al-Ars, Zaid
AU - Berendsen, Floris F.
PY - 2018
Y1 - 2018
N2 - Currently, non-rigid image registration algorithms are too computationally intensive to use in time-critical applications. Existing implementations that focus on speed typically address this by either parallelization on GPU-hardware, or by introducing methodically novel techniques into CPU-oriented algorithms. Stochastic gradient descent (SGD) optimization and variations thereof have proven to drastically reduce the computational burden for CPU-based image registration, but have not been successfully applied in GPU hardware due to its stochastic nature. This paper proposes 1) NiftyRegSGD, a SGD optimization for the GPU-based image registration tool NiftyReg, 2) random chunk sampler, a new random sampling strategy that better utilizes the memory bandwidth of GPU hardware. Experiments have been performed on 3D lung CT data of 19 patients, which compared NiftyRegSGD (with and without random chunk sampler) with CPU-based elastix Fast Adaptive SGD (FASGD) and NiftyReg. The registration runtime was 21.5s, 4.4s and 2.8s for elastix-FASGD, NiftyRegSGD without, and NiftyRegSGD with random chunk sampling, respectively, while similar accuracy was obtained. Our method is publicly available at https://github.com/SuperElastix/NiftyRegSGD.
AB - Currently, non-rigid image registration algorithms are too computationally intensive to use in time-critical applications. Existing implementations that focus on speed typically address this by either parallelization on GPU-hardware, or by introducing methodically novel techniques into CPU-oriented algorithms. Stochastic gradient descent (SGD) optimization and variations thereof have proven to drastically reduce the computational burden for CPU-based image registration, but have not been successfully applied in GPU hardware due to its stochastic nature. This paper proposes 1) NiftyRegSGD, a SGD optimization for the GPU-based image registration tool NiftyReg, 2) random chunk sampler, a new random sampling strategy that better utilizes the memory bandwidth of GPU hardware. Experiments have been performed on 3D lung CT data of 19 patients, which compared NiftyRegSGD (with and without random chunk sampler) with CPU-based elastix Fast Adaptive SGD (FASGD) and NiftyReg. The registration runtime was 21.5s, 4.4s and 2.8s for elastix-FASGD, NiftyRegSGD without, and NiftyRegSGD with random chunk sampling, respectively, while similar accuracy was obtained. Our method is publicly available at https://github.com/SuperElastix/NiftyRegSGD.
KW - GPGPU
KW - memory access optimization
KW - Non-rigid image registration
KW - random chunk sampling
KW - stochastic gradient descent
UR - http://www.scopus.com/inward/record.url?scp=85047351934&partnerID=8YFLogxK
UR - http://resolver.tudelft.nl/uuid:611c4637-6acc-4812-b5f1-8cd66530b0fa
U2 - 10.1117/12.2293098
DO - 10.1117/12.2293098
M3 - Conference contribution
AN - SCOPUS:85047351934
T3 - Proceedings of SPIE
SP - 1
EP - 8
BT - Medical Imaging 2018
A2 - Angelini, Elsa D.
A2 - Landman, Bennett A.
PB - SPIE
CY - Bellingham, WA
T2 - Medical Imaging 2018: Ultrasonic Imaging and Tomography
Y2 - 10 February 2018 through 15 February 2018
ER -