Point-based VR visualization for large-scale mesh datasets by real-time remote computation

  • Authors:
  • Jinghua Ge;Daniel J. Sandin;Andrew Johnson;Tom Peterka;Robert Kooima;Javier I. Girado;Thomas A. DeFanti

  • Affiliations:
  • University of Illinois at Chicago;University of Illinois at Chicago;University of Illinois at Chicago;University of Illinois at Chicago;University of Illinois at Chicago;University of Illinois at Chicago;University of Illinois at Chicago

  • Venue:
  • Proceedings of the 2006 ACM international conference on Virtual reality continuum and its applications
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

High speed interactive visualization of large-scale mesh datasets for desktop VR facilities is still a challenge because of the slow geometry setup and rasterization for huge number of small triangles. This paper presents a point-based virtual reality (VR) visualization pipeline for large-scale mesh datasets in a client-server architecture. Remote server computation which samples the triangle mesh into discrete 2D grids is steered by the client-end interactive frustum request. A point-based geometry is built up incrementally during run time for both server and client. By organizing the point model into a multi-resolution octree-based space partition hierarchy, the client-end visualization ensures fast view reconstruction by splatting the available points onto the screen with efficient occlusion culling and view-dependent level of detail (LOD) control. The combination of the high-priority client side local splatting and server side low-speed view updating decreases the dependence on remote computation performance and network requirements for an interactive VR visualization.