부산대학교 도서관

Technological advances for measuring or simulating volume data have led to large data sizes in many research areas such as biology, medicine, physics, and geoscience. Here, large data can refer to individual data sets with high spatial and/or temporal resolution as well as collections of data sets in the sense of cohorts or ensembles. Therefore, general-purpose and customizable volume visualization and processing systems have to provide out-of-core mechanisms that allow for handling and analyzing such data. Voreen is an open-source rapid-prototyping framework that was originally designed to quickly create custom visualization applications for volumetric imaging data using the meanwhile quite common data flow graph paradigm. In recent years, Voreen has been used in various interdisciplinary research projects with an increasing demand for large data processing capabilities without relying on cluster compute resources. In its latest release, Voreen has thus been extended by out-of-core techniques for processing and visualization of volume data with very high spatial resolution as well as collections of volume data sets including spatio-temporal multi-field simulation ensembles. In this paper we compare state-of-the-art volume processing and visualization systems and conclude that Voreen is the first system combining out-of-core processing and rendering capabilities for large volume data on consumer hardware with features important for interdisciplinary research. We describe how Voreen achieves these goals and show-case its use, performance, and capability to support interdisciplinary research by presenting typical workflows within two large volume data case studies.
Comment: Dominik Drees and Simon Leistikow contributed equally to this work