LBNL Visualization Group Research

Current Projects

High Performance Visualization: Query-Driven Visualization	One of our research thrusts in high performance visualization is known as "query-driven visualization." In this approach, visualization and analysis processing is restricted to a set of data deemed to be "interesting." This approach is an alternative to other techniques that result in ever larger, scalable systems. Application to Hero-Sized Network Traffic Data Analysis. Understanding relationships between variables in complex datasets.
svPerfGL Scientific Visualization OpenGL graphics benchmark	svPerfGL is an OpenGL benchmark intended to measure "real world" performance of scientific visualization applications. These applications are characterized by relatively high payload (i.e., lots of triangles) with relatively few OpenGL state changes. This application takes as input disjoint triangle payload contained in files in netCDF format, renders the frames over a user-specified time duration, rotates the entire scene by one degree per frame, then computes and reports a "triangles per second" performance metric upon exit. More information.
Deep Sky Map	In response to the needs of several astrophysics projects hosted at NERSC, P. Nugent has begun to create an all-sky digital image based upon the point-and-stare observations taken via the Palomar-QUEST Consortium and the SN Factory + Near Earth Asteroid Team. The data spans 7 years and almost 20,000 square degrees, with typically 10-100 pointing on a particular part of the sky. The entire dataset is 60 TB and will create both a temporal and static catalog of astrophysical objects. When completed, the Deep Sky Map will serve as a reference dataset for use by the astrophysical research community. This work was conducted as part of the NERSC Analytics team effort to support a large body of astrophysics researchers who have projects hosted at NERSC. More information.
Acceleratorr SAP & H5Part Development	The Accelerator SAP project aims to develop a simple HDF5 (Hierarchical Data Format) file schema as well as an API that simplifies reading and writing using the HDF5 library suitable to the Accelerator Modelling community. The motivation for this work is to produce a file format that is suitable for large-scale particle simulations. The requirements are the following: it must be machine independent, self-describing, easily extensible, language independent, efficient (serial and parallel) and produces files that are seamlessly shared by different programs. More information
MBender	MBender, or "media-bender", is a research project that focuses on leveraging standard and novel media delivery mechanisms to support interactive, 3D scientific visualization in a remote and distributed context. We leverage QuickTime VR Object movies as a delivery vehicle to support remote, interactive, 3D visualization, and are exploring ways to add multiresolution capability to overcome the fixed-resolution limits of QuickTime VR. More information.
SpectraVis	The current approach for comparing supernovae spectra and light curves is to create an x/y plot that superimposes the two curves, and a "chi by eye" technique is applied to determine whether or not two such datasets are similar. As the number of spectra increase from both observations and simulations, this imprecise approach will not be tractable in the future. A potential solution could be feature detection and classification via machine learning and clustering algorithms applied to the spectral datasets. The end goal is feature detection and similarity detection across supernova spectra. The optimal presentation of these clusters is still an open problem. Another research topic is spectra parameter fitting. About 50 parameters, including minimum and maximum velocity of ionic species, strength or number of ions, and ion temperature, can be said to characterize a supernova spectrum. Software exists which can repeatedly fit experimental data to a model defined by these 50 parameters, and produce the best fit model for each spectrum. More information.
Sunfall	The Visualization Group participated in the development of Sunfall, a collaborative visual analytics system for the Nearby Supernova Factory (http://snfactory.lbl.gov), an international astrophysics experiment and the largest data volume supernova search currently in operation. Sunfall utilizes novel interactive visualization and analysis techniques to facilitate deeper scientific insight into complex, noisy, high-dimensional, high-volume, time-critical data. The system combines novel image processing algorithms, statistical analysis, and machine learning with highly interactive visual interfaces to enable collaborative, user-driven scientific exploration of supernova image and spectral data. Sunfall is currently in operation at the Nearby Supernova Factory; it is the first visual analytics system in production use at a major astrophysics project. More information.
Adaptive Mesh Refinement (AMR) Visualization	AMR data consists of block-structured, hierarchical meshes. AMR is used prolifically by the Applied Numerical Algorithms Group and Center for Computational Science and Engineering at LBNL. AMR is useful for locally increasing resolution in a simulation without incurring the cost of having such increased resolution propogate throughout the entire computational domain. It presents special challenges for scientific visualization due to its multiresolution nature. Our research projects focus on effective ways to perform visualization of AMR data. More information.

Previous Projects

Fast Contour Descriptors for Supernovae	Visualization Group member Cecilia Aragon collaborated on the development of a fast contour descriptor algorithm which was applied to a high-volume supernova detection system for the Nearby Supernova Factory. The shape-detection algorithm reduced the number of false positives generated by the supernova search pipeline by 41% while producing no measurable impact on running time. Because the number of Fourier terms to be calculated is fixed and small, the algorithm runs in linear time, rather than the O(n log n) time of an FFT. More information.
Machine Learning for Supernova Detection	Visualization Group members Raquel Romano and Cecilia Aragon used supervised learning techniques (Support Vector Machines (SVMs), boosted decision trees, random forests) to automatically classify all incoming supernova images for the Nearby Supernova Factory on a nightly basis and rank-order them by the classifier decision value, allowing astrophysicists to quickly examine the 20 or so most promising candidates arriving each morning. More information.
ProteinShop	Funded by Laboratory Directed Research and Development funding, Silvia Crivelli has undertaken an ambitious project to accelerate the challenging problem of predicting and optimizing the shape of protein molecules, also known as "protein folding." Her work has resulted in an interactive 3D visualization application that has increased the size and complexity of protein molecules that can be processed, as shown by the results of her team's performance in the bi-annual CASP competition. More information.
DiVA	The Distributed Visualization Architecture (DiVA) aims to identify and implement a component-based framework for scientific visualization and data analysis. The overall objective of this effort is to promote interoperability within the software tools created by the visualization data analysis research communities, with particular emphasis upon software that runs in remote and distributed, as well as parallel environmnents. More information.
Dex	The Dex project, which is part of the DiVA research effort, aims to combine scientific data management and visualization technology with the objective of improving visual data analysis performance on large and complex scientific datasets. The basic idea is that rather than perform analysis on "the entire dataset" - which is impractical or impossible with large datasets - only a user-specified subset is selected for analysis. The selection criteria consists of a set of boolean queries. More information.
	(1998-2003) Visapult is a pipelined-parallel volume rendering application capable of rendering extremely large volume data on a wide range of common platforms. If you have time to burn and are interested in the historical origins of Visapult, you can read an extensive set of web pages that document early development efforts when Visapult was just undergoing early design and implementation as the Next Generation Internet's reference application. Download Visapult source tarball.
Multiple Gene Sequence Alignment Visualization	More information.
Combustion Visualization (AMR) LDRD	The goal of this project is to evaulate and demonstrate some techniques for visualization specifically for combustion data with embedded boundary conditions. More information. Additional work was performed on this project "on the sly" during FY2000.
The Virtual Protractor	The Virtual Protractor, Summer-Fall 1999 A virtual protractor is used to measure angles between objects perceived in stereo image pairs using virtual reality technology. Each image is generated from a scanning electron microscope, and stereo pairs are obtained by either manipulating the specimen or the electron beam. Images from slightly different viewpoints are combined using stereoscopic rendering to create the illusion of a 3D scene. Virtual reality user interface technology is used to manipulate a virtual measuring device with the goal of accurately measuring spatial characteristics of 3D objects perceived while viewing the stereo image pairs. This approach relies on the human observers' ability to successfully fuse stereo geometry and image-based data. A paper was presented at IEEE Vis99 that describes our work. Download semViewer source tarball.
DataBlaster Toolkit	DataBlaster Toolkit , 1998. The intent of this toolset is to provide the means to easily move data from simulations to visualization tools. Unlike some other code instrumentation tools (such as CUMULVS), there are absolutely no dependancies or restrictions with respect to MP computing environments. The underlying presumption in this toolkit is that the overall goal is to send data, which can be up to a five-dimensional array of double precision floating point values, from a computational source to a consumer. The underlying data must be reducable to a contiguous chunk of memory. This tool makes use of the "eXternal Data Representation" (XDR) libraries for transmission and translation from one architecture to another. Therefore, you can send compute data on an 8-byte-word big-endian machine and consume the data on a 4-byte-word little endian workstation. XDR takes care of all the architecture representation issues (thus, XDR must be present on both client and server machines).
Parametric Visualization and Computation of Large Geochemical Datasets	(1998) This project seeks to make advances in geochemical modeling tools by using computational facilities at Berkeley Lab; seeks to make advances in techniques for visualization of large datasets such as those produced by geochemical simulation models on Berkeley Lab/NERSC equipment; and seeks to make advances in techniques for visualization of divergent data from different sources, some of which is computed and some of which is observed. More information.
Advanced Computational Technology Initiative (ACTI)	The Advanced Computational Technology Initiative (1995-1997). A collaboration with the oil industry produced new commercially available products using visualization tools developed by the Visualization Group at Berkeley Lab.
AVS5 VRModules	How can one bring Virtual Reality to the desktop? Our VRModules Library is a collection of AVS Modules that implement VR. This link describes the contents of the VRModules library. (Circa 1996). More information.
UTCHEM - Coupling VR, Visualization and Simulations	UTCHEM Project A prototype interface - our foray into the world of VR and scientific computing. Fall, 1993.