Yasuhiro Kawasaki

A GPGPU approach for accelerating 2-d/3-d rigid registration of medical images

This paper presents a fast 2-D/3-D rigid registration method using a GPGPU approach, which stands for general-purpose computation on the graphics processing unit (GPU). Our method is based on an intensity-based registration algorithm using biplane images. To accelerate this algorithm, we execute three key procedures of 2-D/3-D registration on the GPU: digitally reconstructed radiograph (DRR) generation, gradient image generation, and normalized cross correlation (NCC) computation. We investigate the usability of our method in terms of registration time and robustness. The experimental results show that our GPU-based method successfully completes a registration task in about 10 seconds, demonstrating shorter registration time than a previous method based on a cluster computing approach.

High-performance computing service over the Internet for intraoperative image processing

This paper presents a framework for a cluster system that is suited for high-resolution image processing over the Internet during surgery. The system realizes high-performance computing (HPC) assisted surgery, which allows surgeons to utilize HPC resources remote from the operating room. One application available in the system is an intraoperative estimator for the range of motion (ROM) adjustment in total hip replacement (THR) surgery. In order to perform this computation-intensive estimation during surgery, we parallelize the ROM estimator on a cluster of 64 PCs, each with two CPUs. Acceleration techniques such as dynamic load balancing and data compression methods are incorporated into the system. The system also provides a remote-access service over the Internet with a secure execution environment. We applied the system to an actual THR surgery performed at Osaka University Hospital and confirmed that it realizes intraoperative ROM estimation without degrading the resolution of images and limiting the area for estimations.

A parallel implementation of 2-D/3-D image registration for computer-assisted surgery

Image registration is a technique usually used for aligning two different images taken at different times and/or from different viewing points. A key challenge for medical image registration is to minimise computation time with a small alignment error in order to realise computer-assisted surgery. In this paper, we present the design and implementation of a parallel two-dimensional/three-dimensional (2-D/3-D) image registration method for computer-assisted surgery. Our method exploits data parallelism and speculative parallelism, aiming at making computation time short enough to carry out registration tasks during surgery. Our experiments show that exploiting both parallelisms reduces computation time on a cluster of 64 PCs from a few tens of minutes to less than a few tens of seconds, a clinically compatible time.

A Parallel Implementation of 2-D/3-D Image Registration for Computer-Assisted Surgery

This paper presents the design and implementation of a parallel two-dimensional/three-dimensional (2-D/3-D) image registration method for computer-assisted surgery. Our method exploits data and speculative parallelism, aiming at making computation time short enough to carry out registration tasks during surgery. Our experiments show that exploiting both parallelisms reduces computation time on a cluster of 64 PCs from a few tens of minutes to less than a few tens of seconds

A high-performance computing service over the Internet for nonrigid image registration

Abstract This paper presents a novel high-performance computing (HPC) system for compute-intensive medical applications. Our system provides surgeons a framework for utilizing remote HPC resources with a secure execution environment by a public key cryptography and a high-speed data transmission by a real-time lossless data compression algorithm. Our implementation on a cluster of 64 off-the-shelf PCs with 128 processors has successfully demonstrated HPC benefits by reducing the turn-around time for the nonrigid registration of liver CT images of 512×512×159 voxels from about 17 h on a sequential system to about 12 min.

Grid Resource Monitoring and Selection for Rapid Turnaround Applications

In this paper, we present a resource monitoring and selection method for rapid turnaround grid applications (for example, within 10 seconds). The novelty of our method is the distributed evaluation of resources for rapidly selecting the appropriate idle resources. We integrate our method with a widely used resource management system, namely the Monitoring and Discovery System 2 (MDS2), and compare our method with the original MDS2 in terms of the performance and the scalability. The performance is measured using a 64-node cluster of PCs and the scalability is analyzed using a theoretical model and the measured performance. The experimental results show that our method reduces the resource selection time by 82%, as compared with the original MDS2. The scalability analysis also indicates that our method can keep the resource selection time within 1 second, up to 500 nodes in local-area-network (LAN) environments. In addition, some simulation results are presented to estimate the impact of our method for wide-area-network (WAN) environments.

Generation of resonance-dependent oscillation by mGluR-I activation switches single spiking to bursting in mesencephalic trigeminal sensory neurons

The primary sensory neurons supplying muscle spindles of jaw‐closing muscles are unique in that they have their somata in the mesencephalic trigeminal nucleus (MTN) in the brainstem, thereby receiving various synaptic inputs. MTN neurons display bursting upon activation of glutamatergic synaptic inputs while they faithfully relay respective impulses arising from peripheral sensory organs. The persistent sodium current (INaP) is reported to be responsible for both the generation of bursts and the relay of impulses. We addressed how INaP is controlled either to trigger bursts or to relay respective impulses as single spikes in MTN neurons. Protein kinase C (PKC) activation enhanced INaP only at low voltages. Spike generation was facilitated by PKC activation at membrane potentials more depolarized than the resting potential. By injection of a ramp current pulse, a burst of spikes was triggered from a depolarized membrane potential whereas its instantaneous spike frequency remained almost constant despite the ramp increases in the current intensity beyond the threshold. A puff application of glutamate preceding the ramp pulse lowered the threshold for evoking bursts by ramp pulses while chelerythrine abolished such effects of glutamate. Dihydroxyphenylglycine, an agonist of mGluR1/5, also caused similar effects, and increased both the frequency and impedance of membrane resonance. Immunohistochemistry revealed that glutamatergic synapses are made onto the stem axons, and that mGluR1/5 and Nav1.6 are co‐localized in the stem axon. Taken together, glutamatergic synaptic inputs onto the stem axon may be able to switch the relaying to the bursting mode.

Real-Time Estimation of Hip Range of Motion for Total Hip Replacement Surgery

This paper presents the design and implementation of a range of motion estimation method that is capable of fine-grained estimation during total hip replacement (THR) surgery. Our method combines an adaptive refinement strategy with a high performance computing system in order to enable real-time estimation. The experimental results indicate that the implementation on a cluster of 64 PCs enables intraoperative estimation of 360 × 360 × 180 stance configurations within a half minute, and thereby plays a key role in selecting and aligning the optimal combination of artificial joint components during THR surgery.

Parallel Adaptive Estimation of Hip Range of Motion for Total Hip Replacement Surgery*A preliminary version of this paper was presented at the 7th Int'l Conf. Medical Image Computing and Computer-Assisted Intervention (MICCAI 2004).

This paper presents the design and implementation of a hip range of motion (ROM) estimation method that is capable of fine-grained estimation during total hip replacement (THR) surgery. Our method is based on two acceleration strategies: (1) adaptive mesh refinement (AMR) for complexity reduction and (2) parallelization for further acceleration. On the assumption that the hip ROM is a single closed region, the AMR strategy reduces the complexity for N × N × N stance configurations from O(N3) to O(ND), where 2 ≤ D ≤ 3 and D is a data-dependent value that can be approximated by 2 in most cases. The parallelization strategy employs the master-worker paradigm with multiple task queues, reducing synchronization between processors with load balancing. The experimental results indicate that the implementation on a cluster of 64 PCs completes estimation of 360 × 360 × 180 stance configurations in 20 seconds, playing a key role in selecting and aligning the optimal combination of artificial joint components during THR surgery.

A High Performance Computing System for Medical Imaging in the Remote Operating Room

This paper presents a novel cluster system, named MI-Cluster, for the purpose of developing a testbed for the cluster-assisted surgery. Our system provides a framework for utilizing high performance computing resources from the remote operating room. One current application is an accurate simulator for the range of motion (ROM) adjustment in total hip replacement (THR) surgery. To perform high-quality imaging during surgery, we have parallelized this compute-intensive ROM simulator on a cluster of PCs with 128 processors. Acceleration techniques such as dynamic load-balancing and data compression have been incorporated into our system. The system also provides a remote access service with a secure execution environment. We applied the system to an actual THR surgery performed at Osaka University Hospital and confirmed that the MI-Cluster system realizes intraoperative ROM simulation without degrading the accuracy of the simulation.