Project

Back to overview

Interactive Multiscale Visualization and Analysis of Very Large Structural Volume Data

Applicant Pajarola Renato
Number 132521
Funding scheme Project funding (Div. I-III)
Research institution Institut für Informatik Universität Zürich
Institution of higher education University of Zurich - ZH
Main discipline Information Technology
Start/End 01.01.2011 - 31.03.2015
Approved amount 269'320.00
Show all

All Disciplines (2)

Discipline
Information Technology
Anthropology, Primatology

Keywords (10)

Scientific Visualization; Structural Volume Rendering; Computer Graphics; Computer Assisted Paleoanthropology; Data Compression; volume rendering; multiscale volume features; tensor approximation; large scale rendering; multiresolution modeling

Lay Summary (English)

Lead
Lay summary
Dense sampling of multidimensional data has become a major trend in fields as diverse as bio-medicine, physics, engineering or geography. Today, data produced with advanced technologies such as micro- or nano-tomography often reaches the size of tens to hundreds of Gigabytes (GB), and exhibit an increasing level of structural complexity. These Very Large Structural Volume (VLSV) data represent a grand challenge for interactive visualization and analysis.This project proposes a new approach to tackle the issues of effective representation of structural volume features as well as efficient interactive rendering, exploration and analysis of VLSV data. The idea is to use a common mathematical framework - tensor approximation - for the three tasks: (1) extraction of relevant structural features at variable scales from VLSV data prior to visualization, (2) feature sensitive data reduction, and (3) interactive multiresolution visualization.The major contributions include: Novel use of an advanced mathematical framework to extract, represent and visualize multiresolution structural volume features in the context of bio-medical imaging. Evaluation of multiscale numerical approximation methods for volume compression and feature extraction. Development of algorithms to efficiently render VLSV data from this novel multiscale basis representation, to extract and visualize structural features as well as to accelerate rendering time. Integration of this structural volume representation and visualization approach into a out-of-core and high-performance parallel rendering system. Implementations of tools to quantify, analyze and visualize microstructures in biological volume data.The proposed project is anticipated to have a significant impact in both research and application of multiscale and parallel volume rendering alike. The targeted VLSV visualization and analysis methods will have applications in various scientific visualization areas, ranging from materials sciences, biomedical imaging and structural biology to physics and data mining. The prime application of the algorithms developed in this project is paleoanthropology. VLSV data sets from hominid fossils contain a large amount of complex structural information, but only a small fraction of the total information can currently be analyzed interactively. This project represents a significant new step toward interactive 3D exploration, visualization and analysis of VLSV data sets. It is expected that the tools developed here will help to shed new light on spatiotemporal patterns of dental growth, and of trabecular remodelling in fossil hominids and modern humans.
Direct link to Lay Summary Last update: 21.02.2013

Responsible applicant and co-applicants

Employees

Publications

Publication
Analysis of Tensor Approximation for Compression-domain Volume Visualization
Ballester-Ripoll Rafael, Suter Susanne K., Pajarola Renato (2015), Analysis of Tensor Approximation for Compression-domain Volume Visualization, in Computers & Graphics, 47, 34-47.
State-of-the-art in Compressed GPU-based Direct Volume Rendering
Balsa Rodriguea Marcos, Gobbetti Enrico, Iglesias Guitan Jose Antonio, Makhinya Maxim, Marton Fabio, Pajarola Renato, Suter Susanne K. (2014), State-of-the-art in Compressed GPU-based Direct Volume Rendering, in Computer Graphics Forum, 33(6), 77-100.
A survey of compressed GPU direct volume rendering
Balsa Rodrìguez Marcos, Gobbetti Enrico, Iglesias Guitiàn José Antonio , Makhinya Maxim, Marton Fabio, Pajarola Renato, Suter Susanne K. (2013), A survey of compressed GPU direct volume rendering, in Eurographics State of The Art Report (STAR).
On Best Rank One Approximation of Tensors
Friedland Shmuel, Mehrmann Volker, Pajarola Renato, Suter Susanne K. (2013), On Best Rank One Approximation of Tensors, in Numerical Linear Algebra with Applications, 20(6), 942-955.
TAMRESH: Tensor Approximation Multiresolution Hierarchy for Interactive Volume Visualization
Suter Susanne K., Makhinya Maxim, Pajarola Renato (2013), TAMRESH: Tensor Approximation Multiresolution Hierarchy for Interactive Volume Visualization, in Computer Graphics Forum, 32(3), 151-160.
Tensor Approximation in Visualization and Graphics
Pajarola Renato, Suter Susanne K., Ruiters Roland (2013), Tensor Approximation in Visualization and Graphics, in Eurographics Tutorials.
Guest Editor's Introduction: Special Section on the Eurographics Symposium on Parallel Graphics and Visualization (EGPGV)
Guest Editor's Introduction: Special Section on the Eurographics Symposium on Parallel Graphics and Visualization (EGPGV), (2012), Guest Editor's Introduction: Special Section on the Eurographics Symposium on Parallel Graphics and Visualization (EGPGV), 18(6).
Information Overload
Pajarola Renato (2012), Information Overload, in Public Service Review: European Science & Technology, (16), 76-77.
Tensor Approximation Properties for Multiresolution and Multiscale Volume Visualization
Suter Susanne K., Pajarola Renato (2012), Tensor Approximation Properties for Multiresolution and Multiscale Volume Visualization, in Posters IEEE Visualization Conference.
Interactive Multiscale Tensor Reconstruction for Multiresolution Volume Visualization
Suter Susanne K., Iglesias Guitian Jose Antonio, Marton Fabio, Agus Marco, Elsener Andreas, Zollikofer Christoph P.E., Gopi M., Gobbetti Enrico, Pajarola Renato (2011), Interactive Multiscale Tensor Reconstruction for Multiresolution Volume Visualization, in IEEE Transactions on Visualization and Computer Graphics, 17(12), 2135-2143.
Proceedings Symposium on Parallel Graphics and Visualization
(2011), Proceedings Symposium on Parallel Graphics and Visualization, Eurographics Association, -.

Collaboration

Group / person Country
Types of collaboration
TU Berlin Germany (Europe)
- Publication
University of California Irvine United States of America (North America)
- in-depth/constructive exchanges on approaches, methods or results
- Publication
- Exchange of personnel
University of Illinois Chicago United States of America (North America)
- in-depth/constructive exchanges on approaches, methods or results
- Publication
EPFL, Lausanne Switzerland (Europe)
- in-depth/constructive exchanges on approaches, methods or results
- Publication
- Research Infrastructure
CRS4, Sardegna Ricerche Italy (Europe)
- in-depth/constructive exchanges on approaches, methods or results
- Publication
- Exchange of personnel

Scientific events

Active participation

Title Type of contribution Title of article or contribution Date Place Persons involved
Dagstuhl Seminar 14082 on Visualization and Processing of Higher Order Descriptors for Multi-Valued Data Talk given at a conference Multiscale Volume Visualization 16.02.2014 Schloss Dagstuhl, Germany Pajarola Renato;
Eurographics Visualization Conference Talk given at a conference TAMRESH - Tensor Approximation Multiresolution Hierarchy for Interactive Volume Visualization 19.06.2013 Leipzig, Germany Suter Susanne;
Eurographics Conference State-of-the-art Presentations Talk given at a conference A Survey of Compressed GPU Direct Volume Rendering 09.05.2013 Girona, Spain Suter Susanne; Pajarola Renato;
Eurographics Conference Tutorials Talk given at a conference Tensor Approximation in Visualization and Graphics 07.05.2013 Girona, Spain Pajarola Renato; Suter Susanne;
KAUST Visualization Summit Individual talk Tensor Approximation in Scientific Visualization 16.04.2013 Jeddha, Saudi Arabia Pajarola Renato;
IEEE Visualization Conference Poster Tensor Approximation Properties for Multiresolution and Multiscale Volume Visualization 14.10.2012 Seattle, USA, United States of America Pajarola Renato; Suter Susanne;
Annual CIMST Meeting Poster Multiresolution and Multiscale Volume Visualization Using Tensor Approximation 03.09.2012 Zürich, Switzerland Suter Susanne;
Mini-Symposium Imaging and Visualization Talk given at a conference Data Intensive Interactive Scientific Visualization 21.11.2011 Zürich, Switzerland Pajarola Renato;
IEEE Visualization Conference Talk given at a conference Interactive multiscale tensor reconstruction for multiresolution volume visualization 23.10.2011 Providence, United States, United States of America Pajarola Renato; Suter Susanne;


Self-organised

Title Date Place
Eurographics Symposium on Parallel Graphics and Visualization 14.05.2012 Cagliari, Italy, Italy
Eurographics Symposium on Parallel Graphics and Visualization 12.04.2011 Llandudno, United Kingdom, Great Britain and Northern Ireland

Communication with the public

Communication Title Media Place Year
Talks/events/exhibitions Data intensive interactive visualization and analysis challenges International 2012
Talks/events/exhibitions Ein Tag im Leben einer Informatikerin German-speaking Switzerland 2011

Awards

Title Year
Jahrespreis der Universität Zürich 2013

Associated projects

Number Title Start Funding scheme
121072 High-Performance Visualization 01.04.2008 International short research visits
116329 VISS: Visualization Infrastructures for Scalable Scientific Computing 01.04.2007 Project funding (Div. I-III)

Abstract

Modern data acquisition and sampling technologies in natural sciences and engineering yield data sets in the order of magnitude of tens to hundreds of Gigabytes, very often encoding highly complex spatial structures of tissues or materials that are sampled. These Very Large Structural Volume (VLSV) data sets represent grand challenges for interactive visualization and analysis, because the overall amount of data to be processed is typically beyond the limits of current computer and graphics hardware performance. This project proposes a unified approach to resolve this bottleneck. It employs higher-order tensor approximation methods as a common mathematical framework to extract relevant features from VLSV (thus reducing the amount of data), to further compress these data, and to analyse them visually using interactive rendering. This new multiscale feature-preserving compression and visualization approach is to be integrated into a multiresolution parallel rendering framework. As a result, a user interface and tools for interactive exploration and analysis of VLSV data are provided. For verification of the developed concepts, it uses 10-100GB VLSV data acquired with microtomographic methods from teeth and bones of fossil and modern humans. The aim is to visualize the 3-dimensional texture of daily incremental growth structures in teeth, and to visualize structural and biomechanical properties of trabecular bone from small (local) to large (global) scales.
-