An Object Selecting Method for Accelerating Volume Rendering of Large Datasets

N. Fallah and A. Eydgahi (USA)

Keywords

Visualization, volume rendering, object selecting, large datasets, multi-resolution, and simulation.

Abstract

This paper presents a new technique for selecting proper objects for accelerating visualization of large datasets similar to urban areas. The proposed method improves the current algorithms by introducing a new technique which uses statistics about different objects to cover more general and comprehensive datasets. In this technique, proximity, level of details, and statistics from previous moves are used to predict next position and to select objects for future use. A multi-resolution approach is introduced which is based on a special shape. By gathering statistics about different objects a unique shape called layer is obtained that is used to select different objects and put them in the groups to be rendered with different resolutions. This way acceleration to produce high resolution visualization at interactive rate is achieved. The proposed technique solves the issues of redundant object downloading and redundant rendering on the server by downloading objects and images based on software and hardware specifications. This way it saves server and network resources to reduce the delay and to achieve interactive frame rate on client side.

Important Links:



Go Back


IASTED
Rotating Call For Paper Image