BYR Achieve · 镜像论坛

Cooperative Volume Exploration Based on Differential Topology 报告人: Professor Issei Fujishiro, Tohoku University, Japan 报告时间: 2008年10月28日星期二14:00~15:00 报告地点: 清华大学FIT 楼1区415房间主办单位: 清华大学软件学院计算机图形学与辅助设计研究所 Professor Issei Fujishiro is currently a professor of Institute of Fluid Science and Graduate School of Information Sciences at Tohoku University, Sendai, Japan, where he directs Integrated Visual Informatics Laboratory. His research interests include volume visualization and graphics, visualization design and lifecycle management, and multi-modal mixed reality environments. He has served on the editorial boards for Transactions of Information Processing Society of Japan (1997-2001), IEEE Transactions on Visualization and Computer Graphics (1999-2003), International Journal of Image and Graphics (2001-2004), and Computers & Graphics (2003 to date). He was a Program Chair for EG/IEEE Volume Graphics 2003/2005 and IEEE PacificVis 2008; and chaired many international conferences, including IEEE SMI2006 and AIM@SHAPE Shapes and Semantics 2006. Fujishiro is a member of IEEE Computer Society, ACM, SIGGRAPH/SIGMOD, Eurographics, IPSJ, IIEEJ, VSJ, VRSJ, JSCES and JSME. Rapid increase in the performance of computing and measurement environments as well as in opportunities to use the global network infrastructure has given rise to so-called "data big-bang". It has motivated us to develop a cooperative visulization environment which allows the users to deal with a huge amount of data with the aid of data-centric, semi-autormatic parameter tweaking. In the past decade, we have consistently built on differential topology to come up with a set of tools for examining the global structure and local features of a target 4D volume effectively. The topological skeleton of a given snapshot volume can be represented by Volume Skeleton Tree (VST), which is a kind of level-set graph, consisting of volumetric critical points and their global connectivity to delineate the topological evolution, such as appearance, disappearance, splitting, and merging, of an isosurface according to the scalar field value. We have developed an efficient and robust method for constructing proper VSTs even from a complex and/or noisy volume using adaptive tetrahedralization and topological skeleton simplification. Our current tools to explore volumes visually with reference to the extracted VSTs include critical point-based colormap/transfer function generators; critical/representative isosurface and cross-section extractors; optimal viewpoint/light sournce locator; topologically-accentuated multi-dimensional transfer function designer; interval volume decomposer; and T-Map (critical timing finder and selective data migration). In this talk, the effectiveness of these cooperative volume exoloration tools will be illustrated with applications to real datasets from many disciplines.

[学术讲座]Cooperative Volume Exploration Based on Different