Dataset Information

Efficient methods for implementation of multi-level nonrigid mass-preserving image registration on GPUs and multi-threaded CPUs.

ABSTRACT: BACKGROUND AND OBJECTIVE:Faster and more accurate methods for registration of images are important for research involved in conducting population-based studies that utilize medical imaging, as well as improvements for use in clinical applications. We present a novel computation- and memory-efficient multi-level method on graphics processing units (GPU) for performing registration of two computed tomography (CT) volumetric lung images. METHODS:We developed a computation- and memory-efficient Diffeomorphic Multi-level B-Spline Transform Composite (DMTC) method to implement nonrigid mass-preserving registration of two CT lung images on GPU. The framework consists of a hierarchy of B-Spline control grids of increasing resolution. A similarity criterion known as the sum of squared tissue volume difference (SSTVD) was adopted to preserve lung tissue mass. The use of SSTVD consists of the calculation of the tissue volume, the Jacobian, and their derivatives, which makes its implementation on GPU challenging due to memory constraints. The use of the DMTC method enabled reduced computation and memory storage of variables with minimal communication between GPU and Central Processing Unit (CPU) due to ability to pre-compute values. The method was assessed on six healthy human subjects. RESULTS:Resultant GPU-generated displacement fields were compared against the previously validated CPU counterpart fields, showing good agreement with an average normalized root mean square error (nRMS) of 0.044±0.015. Runtime and performance speedup are compared between single-threaded CPU, multi-threaded CPU, and GPU algorithms. Best performance speedup occurs at the highest resolution in the GPU implementation for the SSTVD cost and cost gradient computations, with a speedup of 112 times that of the single-threaded CPU version and 11 times over the twelve-threaded version when considering average time per iteration using a Nvidia Tesla K20X GPU. CONCLUSIONS:The proposed GPU-based DMTC method outperforms its multi-threaded CPU version in terms of runtime. Total registration time reduced runtime to 2.9min on the GPU version, compared to 12.8min on twelve-threaded CPU version and 112.5min on a single-threaded CPU. Furthermore, the GPU implementation discussed in this work can be adapted for use of other cost functions that require calculation of the first derivatives.

SUBMITTER: Ellingwood ND

PROVIDER: S-EPMC4803628 | biostudies-literature | 2016 Apr

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Efficient methods for implementation of multi-level nonrigid mass-preserving image registration on GPUs and multi-threaded CPUs.

Ellingwood Nathan D ND Yin Youbing Y Smith Matthew M Lin Ching-Long CL

Computer methods and programs in biomedicine 20160106

<h4>Background and objective</h4>Faster and more accurate methods for registration of images are important for research involved in conducting population-based studies that utilize medical imaging, as well as improvements for use in clinical applications. We present a novel computation- and memory-efficient multi-level method on graphics processing units (GPU) for performing registration of two computed tomography (CT) volumetric lung images.<h4>Methods</h4>We developed a computation- and memory ...[more]

PMID: 26776541

Similar Datasets

Project description:BackgroundGene Set Enrichment Analysis (GSEA) is a popular method to reveal significant dependencies between predefined sets of gene symbols and observed phenotypes by evaluating the deviation of gene expression values between cases and controls. An established measure of inter-class deviation, the enrichment score, is usually computed using a weighted running sum statistic over the whole set of gene symbols. Due to the lack of analytic expressions the significance of enrichment scores is determined using a non-parametric estimation of their null distribution by permuting the phenotype labels of the probed patients. Accordingly, GSEA is a time-consuming task due to the large number of required permutations to accurately estimate the nominal p-value - a circumstance that is even more pronounced during multiple hypothesis testing since its estimate is lower-bounded by the inverse number of samples in permutation space.ResultsWe present rapidGSEA - a software suite consisting of two tools for facilitating permutation-based GSEA: cudaGSEA and ompGSEA. cudaGSEA is a CUDA-accelerated tool using fine-grained parallelization schemes on massively parallel architectures while ompGSEA is a coarse-grained multi-threaded tool for multi-core CPUs. Nominal p-value estimation of 4,725 gene sets on a data set consisting of 20,639 unique gene symbols and 200 patients (183 cases + 17 controls) each probing one million permutations takes 19 hours on a Xeon CPU and less than one hour on a GeForce Titan X GPU while the established GSEA tool from the Broad Institute (broadGSEA) takes roughly 13 days.ConclusioncudaGSEA outperforms broadGSEA by around two orders-of-magnitude on a single Tesla K40c or GeForce Titan X GPU. ompGSEA provides around one order-of-magnitude speedup to broadGSEA on a standard Xeon CPU. The rapidGSEA suite is open-source software and can be downloaded at https://github.com/gravitino/cudaGSEA as standalone application or package for the R framework.

Project description:The primary aim of this study is to investigate the performance difference of rigid and nonrigid registration schemes in matching corresponding pulmonary nodules depicted on sequential chest computed tomography (CT) examinations.A gradient descent based rigid registration algorithm with scaling was developed and it handled the involved geometric transformations (i.e., translation, rescaling, shearing, and rotation) separately instead of optimizing them in a single pass. Given two lung CT examinations, the scaling and translation parameters were simply estimated from the lung volume dimensions (e.g., size and mass center), while the rotation parameters were optimized progressively using gradient descent. To investigate the performance difference of rigid and nonrigid schemes in pulmonary nodule registration, the well-known nonrigid Demons algorithm was implemented and tested along with the developed schemes against 60 diverse low-dose clinical lung CT examinations with average 2-yr follow-up scans. A verified cancer and its correspondence in the follow-up scan as well as their spatial locations (mass center) were identified in each examination. In addition to the computational efficiency, the accuracy of these registration procedures was assessed by computing the Euclidean distances between the corresponding nodules after the registration. To demonstrate the advantage of the developed algorithm, the authors also implemented a fast iterative closest point (ICP) based rigid algorithm and compared their performance.Our experiments on the collected chest CT examinations showed that the nodule registration errors in 3D Euclidean distance for the developed rigid affine approach, the traditional ICP algorithm, and the refining nonrigid Demons algorithm were 9.6, 9.8, and 10.0 mm, respectively, and the corresponding computational costs in time were 5, 300, and 55 s, respectively.A rigid solution may be preferred in practice for the pulmonary nodule registration in longitudinal studies because of its relatively high efficiency and sufficient accuracy for the clinical need.

Project description:This study introduces a probabilistic nonrigid registration method for use in image-guided prostate brachytherapy. Intraoperative imaging for prostate procedures, usually transrectal ultrasound (TRUS), is typically inferior to diagnostic-quality imaging of the pelvis such as endorectal magnetic resonance imaging (MRI). MR images contain superior detail of the prostate boundaries and provide substructure features not otherwise visible. Previous efforts to register diagnostic prostate images with the intraoperative coordinate system have been deterministic and did not offer a measure of the registration uncertainty. The authors developed a Bayesian registration method to estimate the posterior distribution on deformations and provide a case-specific measure of the associated registration uncertainty.The authors adapted a biomechanical-based probabilistic nonrigid method to register diagnostic to intraoperative images by aligning a physician's segmentations of the prostate in the two images. The posterior distribution was characterized with a Markov Chain Monte Carlo method; the maximum a posteriori deformation and the associated uncertainty were estimated from the collection of deformation samples drawn from the posterior distribution. The authors validated the registration method using a dataset created from ten patients with MRI-guided prostate biopsies who had both diagnostic and intraprocedural 3 Tesla MRI scans. The accuracy and precision of the estimated posterior distribution on deformations were evaluated from two predictive distance distributions: between the deformed central zone-peripheral zone (CZ-PZ) interface and the physician-labeled interface, and based on physician-defined landmarks. Geometric margins on the registration of the prostate's peripheral zone were determined from the posterior predictive distance to the CZ-PZ interface separately for the base, mid-gland, and apical regions of the prostate.The authors observed variation in the shape and volume of the segmented prostate in diagnostic and intraprocedural images. The probabilistic method allowed us to convey registration results in terms of posterior distributions, with the dispersion providing a patient-specific estimate of the registration uncertainty. The median of the predictive distance distribution between the deformed prostate boundary and the segmented boundary was ≤3 mm (95th percentiles within ±4 mm) for all ten patients. The accuracy and precision of the internal deformation was evaluated by comparing the posterior predictive distance distribution for the CZ-PZ interface for each patient, with the median distance ranging from -0.6 to 2.4 mm. Posterior predictive distances between naturally occurring landmarks showed registration errors of ≤5 mm in any direction. The uncertainty was not a global measure, but instead was local and varied throughout the registration region. Registration uncertainties were largest in the apical region of the prostate.Using a Bayesian nonrigid registration method, the authors determined the posterior distribution on deformations between diagnostic and intraprocedural MR images and quantified the uncertainty in the registration results. The feasibility of this approach was tested and results were positive. The probabilistic framework allows us to evaluate both patient-specific and location-specific estimates of the uncertainty in the registration result. Although the framework was tested on MR-guided procedures, the preliminary results suggest that it may be applied to TRUS-guided procedures as well, where the addition of diagnostic MR information may have a larger impact on target definition and clinical guidance.

Dataset Information

Efficient methods for implementation of multi-level nonrigid mass-preserving image registration on GPUs and multi-threaded CPUs.

Publications

Efficient methods for implementation of multi-level nonrigid mass-preserving image registration on GPUs and multi-threaded CPUs.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets