SCIENTIFIC COMPUTING AND IMAGING INSTITUTE
at the University of Utah

An effective approach to structure learning and parameter estimation for Gaussian graphical models is to impose a sparsity prior, such as a Laplace prior, on the entries of the precision matrix. Such an approach involves a hyperparameter that must be tuned to control the amount of sparsity. In this paper, we introduce a parameter-free method for estimating a precision matrix with sparsity that adapts to the data automatically. We achieve this by formulating a hierarchical Bayesian model of the precision matrix with a noninformative Jeffreys' hyperprior. We also naturally enforce the symmetry and positivede definiteness constraints on the precision matrix by parameterizing it with the Cholesky decomposition. Experiments on simulated and real (cell signaling) data demonstrate that the proposed approach not only automatically adapts the sparsity of the model, but it also results in improved estimates of the precision matrix compared to the Laplace prior model with sparsity parameter chosen by cross-validation.

This paper presents a generative Bayesian model for diffeomorphic image registration and atlas building. We develop an atlas estimation procedure that simultaneously estimates the parameters controlling the smoothness of the diffeomorphic transformations. To achieve this, we introduce a Monte Carlo Expectation Maximization algorithm, where the expectation step is approximated via Hamiltonian Monte Carlo sampling on the manifold of diffeomorphisms. An added benefit of this stochastic approach is that it can successfully solve difficult registration problems involving large deformations, where direct geodesic optimization fails. Using synthetic data generated from the forward model with known parameters, we demonstrate the ability of our model to successfully recover the atlas and regularization parameters. We also demonstrate the effectiveness of the proposed method in the atlas estimation problem for 3D brain images.

Principal geodesic analysis (PGA) is a generalization of principal component analysis (PCA) for dimensionality reduction of data on a Riemannian manifold. Currently PGA is defined as a geometric fit to the data, rather than as a probabilistic model. Inspired by probabilistic PCA, we present a latent variable model for PGA that provides a probabilistic framework for factor analysis on manifolds. To compute maximum likelihood estimates of the parameters in our model, we develop a Monte Carlo Expectation Maximization algorithm, where the expectation is approximated by Hamiltonian Monte Carlo sampling of the latent variables. We demonstrate the ability of our method to recover the ground truth parameters in simulated sphere data, as well as its effectiveness in analyzing shape variability of a corpus callosum data set from human brain images.

Multivariate volumetric datasets are important to both science and medicine. We propose a transfer function (TF) design approach based on user selected samples in the spatial domain to make multivariate volumetric data visualization more accessible for domain users. Specifically, the user starts the visualization by probing features of interest on slices and the data values are instantly queried by user selection. The queried sample values are then used to automatically and robustly generate high dimensional transfer functions (HDTFs) via kernel density estimation (KDE). Alternatively, 2D Gaussian TFs can be automatically generated in the dimensionality reduced space using these samples. With the extracted features rendered in the volume rendering view, the user can further refine these features using segmentation brushes. Interactivity is achieved in our system and different views are tightly linked. Use cases show that our system has been successfully applied for simulation and complicated seismic data sets.

We present a volume visualization method that allows interactive rendering and efficient querying of large multivariate seismic volume data on consumer level PCs. The volume rendering pipeline utilizes a virtual memory structure that supports out-of-core multivariate multi-resolution data and a GPU-based ray caster that allows interactive multivariate transfer function design. A Gaussian mixture model representation is precomputed and nearly interactive querying is achieved by testing the Gaussian functions against user defined transfer functions on the GPU in the runtime. Finally, the method has been tested on a multivariate 3D seismic dataset which is larger than the size of the main memory of the testing machine.

Multi-compartment models in diffusion MRI (dMRI) are used to describe complex white matter fiber architecture of the brain. In this paper, we propose a novel multi-compartment estimation method based on the ball-and-stick model, which is composed of an isotropic diffusion compartment (\"ball\") as well as one or more perfectly linear diffusion compartments (\"sticks\"). To model the noise distribution intrinsic to dMRI measurements, we introduce a Rician likelihood term and estimate the model parameters by means of an Expectation Maximization (EM) algorithm. This paper also addresses the problem of selecting the number of fiber compartments that best fit the data, by introducing a sparsity prior on the volume mixing fractions. This term provides automatic model selection and enables us to discriminate different fiber populations. When applied to simulated data, our method provides accurate estimates of the fiber orientations, diffusivities, and number of compartments, even at low SNR, and outperforms similar methods that rely on a Gaussian noise distribution assumption. We also apply our method to in vivo brain data and show that it can successfully capture complex fiber structures that match the known anatomy.

The problem of packing circles into a domain of prescribed topology is considered. The circles need not have equal radii. The Collins-Stephenson algorithm computes such a circle packing. This algorithm is parallelized in two different ways and its performance is reported for a triangular, planar domain test case. The implementation uses the highly parallel graphics processing unit (GPU) on commodity hardware. The speedups so achieved are discussed based on a number of experiments.

Keywords: Circle packing, Algorithm performance, Parallel computation, Graphics processing unit (GPU)

Atrial Fibrosis and Sinus Node Dysfunction. Introduction: Sinus node dysfunction (SND) commonly manifests with atrial arrhythmias alternating with sinus pauses and sinus bradycardia. The underlying process is thought to be because of atrial fibrosis. We assessed the value of atrial fibrosis, quantified using Late Gadolinium Enhanced-MRI (LGE-MRI), in predicting significant SND requiring pacemaker implant.

Methods: Three hundred forty-four patients with atrial fibrillation (AF) presenting for catheter ablation underwent LGE-MRI. Left atrial (LA) fibrosis was quantified in all patients and right atrial (RA) fibrosis in 134 patients. All patients underwent catheter ablation with pulmonary vein isolation with posterior wall and septal debulking. Patients were followed prospectively for 329 ± 245 days. Ambulatory monitoring was instituted every 3 months. Symptomatic pauses and bradycardia were treated with pacemaker implantation per published guidelines.

Results: The average patient age was 65 ± 12 years. The average wall fibrosis was 16.7 ± 11.1% in the LA, and 5.3 ± 6.4% in the RA. RA fibrosis was correlated with LA fibrosis (R²= 0.26; P < 0.01). Patients were divided into 4 stages of LA fibrosis (Utah I: 35%). Twenty-two patients (mean atrial fibrosis, 23.9%) required pacemaker implantation during follow-up. Univariate and multivariate analysis identified LA fibrosis stage (OR, 2.2) as a significant predictor for pacemaker implantation with an area under the curve of 0.704.

Conclusions: In patients with AF presenting for catheter ablation, LGE-MRI quantification of atrial fibrosis demonstrates preferential LA involvement. Significant atrial fibrosis is associated with clinically significant SND requiring pacemaker implantation. (J Cardiovasc Electrophysiol, Vol. 23, pp. 44-50, January 2012)

A finite element formulation of neutral solute transport across a contact interface between deformable porous media is implemented and validated against analytical solutions. By reducing the integral statements of external virtual work on the two contacting surfaces into a single contact integral, the algorithm automatically enforces continuity of solute molar flux across the contact interface, whereas continuity of the effective solute concentration (a measure of the solute mechano-chemical potential) is achieved using a penalty method. This novel formulation facilitates the analysis of problems in biomechanics where the transport of metabolites across contact interfaces of deformable tissues may be of interest. This contact algorithm is the first to address solute transport across deformable interfaces, and is made available in the public domain, open-source finite element code FEBio (http://www.febio.org).

Keywords: FEBio, Finite element modeling, Contact mechanics, Solute transport, Porous media, Biphasic theory

This paper proposes a novel formulation to model and analyze the statistical characteristics of some types of segmentation problems that are based on combining label maps / templates / atlases. Such segmentation-by-example approaches are quite powerful on their own for several clinical applications and they provide prior information, through spatial context, when combined with intensity-based segmentation methods. The proposed formulation models a class of multiatlas segmentation problems as nonparametric regression problems in the high-dimensional space of images. The paper presents a systematic analysis of the nonparametric estimation's convergence behavior (i.e. characterizing segmentation error as a function of the size of the multiatlas database) and shows that it has a specific analytic form involving several parameters that are fundamental to the specific segmentation problem (i.e. chosen anatomical structure, imaging modality, registration method, label-fusion algorithm, etc.). We describe how to estimate these parameters and show that several brain anatomical structures exhibit the trends determined analytically. The proposed framework also provides per-voxel confidence measures for the segmentation. We show that the segmentation error for large database sizes can be predicted using small-sized databases. Thus, small databases can be exploited to predict the database sizes required (\"how many templates\") to achieve \"good\" segmentations having errors lower than a specified tolerance. Such cost-benefit analysis is crucial for designing and deploying multiatlas segmentation systems.

Background: The medial distal tibial angle (MDTA) is used to determine ankle alignment. The mortise view is the standard to measure MDTA, but the hindfoot alignment view (HAV) has become popular. The MDTA may vary between views, influencing the choice of surgery.

Methods: The MDTA was compared between the mortise and HAV in 146 ankles. MDTA was correlated to age and sagittal tibial tilt for each view. Differences in MDTA by gender and ethnicity were assessed. Diagnostic agreement (varus, valgus, normal) between views was calculated. Clinical assessment of alignment was determined and percent agreement between clinical and radiographic alignment was quantified.

Results: The MDTA measured from the mortise view and HAV radiographs was 89.0 (range, 81 to 96 degrees; SD = 2.8) degrees and 86.0 (range, 73 to 95 degrees; SD = 3.5) degrees, respectively. The MDTA was comparable for both genders for mortise (p = 0.356) and HAV (p = 0.621). The MDTA was comparable in all ethnic groups for mortise view (p = 0.616) and HAV (p = 0.916). Correlation between the measured MDTA and age was not statistically significant for both the mortise (r = 0.118; p = 0.158) and HAV (r = 0.148; p = 0.074). In only 47.3% of all ankles was the radiographic diagnosis of alignment the same between views. Agreement between clinical and radiographic classifications was 60.3% for the mortise view and 52.8% for the HAV.

Conclusion: Substantial disagreement in primary alignment was found between the mortise and HAV as quantified by the MDTA. Agreement between clinical and radiographic alignment was also poor. Clinical Relevance: Advanced imaging such as CT or MRI may better describe ankle alignment.

We introduce several useful utilities in development for the creation and analysis of real wildland fire simulations using WRF and SFIRE. These utilities exist as standalone programs and scripts as well as extensions to other well known software. Python web scrapers automate the process of downloading and preprocessing atmospheric and surface data from common sources. Other scripts simplify the domain setup by creating parameter files automatically. Integration with Google Earth allows users to explore the simulation in a 3D environment along with real surface imagery. Postprocessing scripts provide the user with a number of output data formats compatible with many commonly used visualization suites allowing for the creation of high quality 3D renderings. As a whole, these improvements build toward a unified web application that brings a sophisticated wildland fire modeling environment to scientists and users alike.

With the onset of extreme-scale computing, I/O constraints make it increasingly difficult for scientists to save a sufficient amount of raw simulation data to persistent storage. One potential solution is to change the data analysis pipeline from a post-process centric to a concurrent approach based on either in-situ or in-transit processing. In this context computations are considered in-situ if they utilize the primary compute resources, while in-transit processing refers to offloading computations to a set of secondary resources using asynchronous data transfers. In this paper we explore the design and implementation of three common analysis techniques typically performed on large-scale scientific simulations: topological analysis, descriptive statistics, and visualization. We summarize algorithmic developments, describe a resource scheduling system to coordinate the execution of various analysis workflows, and discuss our implementation using the DataSpaces and ADIOS frameworks that support efficient data movement between in-situ and in-transit computations. We demonstrate the efficiency of our lightweight, flexible framework by deploying it on the Jaguar XK6 to analyze data generated by S3D, a massively parallel turbulent combustion code. Our framework allows scientists dealing with the data deluge at extreme scale to perform analyses at increased temporal resolutions, mitigate I/O costs, and significantly improve the time to insight.

This report provides a summary of the status of the Uintah Computation Framework (UCF) software. Uintah is uniquely equipped to tackle large-scale multi-physics science and engineering problems on disparate length and time scales. The Uintah framework makes it possible to run adaptive computations on modern HPC architectures with tens and now hundreds of thousands of cores with complex communication/memory hierarchies. Uintah was orignally developed in the University of Utah Center for Simulation of Accidental Fires and Explosions (C-SAFE), a DOE-funded academic alliance project and then extended to the broader NSF snd DOE science and engineering communities. As Uintah is applicable to a wide range of engineering problems that involve fl uid-structure interactions with highly deformable structures it is used for a number of NSF-funded and DOE engineering projects. In this report the Uintah framework software is outlined and typical applications are illustrated. Uintah is open-source software that is available through the MIT open-source license at http://www.uintah.utah.edu/.

The task-based approach to software and parallelism is well-known and has been proposed as a potential candidate, named the silver model, for exascale software. This approach is not yet widely used in the large-scale multi-core parallel computing of complex systems of partial differential equations. After surveying task-based approaches we investigate how well the Uintah software and an extension named Wasatch fit in the task-based paradigm and how well they perform on large scale parallel computers. The conclusion is that these approaches show great promise for petascale but that considerable algorithmic challenges remain.

Keywords: DOD, Uintah, CSAFE

The placement of tasks in a parallel application on specific nodes of a supercomputer can significantly impact performance. Traditionally, this task mapping has focused on reducing the distance between communicating tasks on the physical network. This minimizes the number of hops that point-to-point messages travel and thus reduces link sharing between messages and contention. However, for applications that use collectives over sub-communicators, this heuristic may not be optimal. Many collectives can benefit from an increase in bandwidth even at the cost of an increase in hop count, especially when sending large messages. For example, placing communicating tasks in a cube configuration rather than a plane or a line on a torus network increases the number of possible paths messages might take. This increases the available bandwidth which can lead to significant performance gains.

We have developed Rubik, a tool that provides a simple and intuitive interface to create a wide variety of mappings for structured communication patterns. Rubik supports a number of elementary operations such as splits, tilts, or shifts, that can be combined into a large number of unique patterns. Each operation can be applied to disjoint groups of processes involved in collectives to increase the effective bandwidth. We demonstrate the use of Rubik for improving performance of two parallel codes, pF3D and Qbox, which use collectives over sub-communicators.

Ray tracing in scientific visualization allows for substantial gains in performance and rendering quality with large scale polygonal datasets compared to brute-force rasterization, however implementing new rendering architectures into existing tools is often costly and time consuming. This paper presents a library, GLuRay, which intercepts OpenGL calls from many common visualization applications and renders them with the CPU ray tracer Manta without modification to the underlying visualization tool. Rendering polygonal models such as isosurfaces can be done identically to an OpenGL implementation using provided material and camera properties or superior rendering can be achieved using enhanced settings such as dielectric materials or pinhole cameras with depth of field effects. Comparative benchmarks were conducted on the Texas Advanced Computing Center’s Longhorn cluster using the popular visualization packages ParaView, VisIt, Ensight, and VAPOR. Through the parallel ren- dering package ParaView, scaling up to 64 nodes is demonstrated. With our tests we show that using OpenGL interception to accelerate and enhance visualization programs provides a viable enhancement to existing tools with little overhead and no code modification while allowing for the creation of publication quality renderings using advanced effects and greatly improved large-scale software rendering performance within tools that scientists are currently using.

Keywords: kaust, scidac

Large-scale analysis and visualization is becoming increasingly important as supercomputers and their simulations produce larger and larger data. These large data sizes are pushing the limits of traditional rendering algorithms and tools thus motivating a study exploring these limits and their possible resolutions through alternative rendering algorithms . In order to better understand real-world performance with large data, this paper presents a detailed timing study on a large cluster with the widely used visualization tools ParaView and VisIt. The software ray tracer Manta was integrated into these programs in order to show that improved performance could be attained with software ray tracing on a distributed memory, GPU enabled, parallel visualization resource. Using the Texas Advanced Computing Center’s Longhorn cluster which has multi-core CPUs and GPUs with large-scale polygonal data, we find multi-core CPU ray tracing to be significantly faster than both software rasterization and hardware-accelerated rasterization in existing scientific visualization tools with large data.

Keywords: kaust, scidac