Theses and Dissertations

Issuing Body

Mississippi State University

Advisor

Jamil, Hasan

Committee Member

Farias, Ricardo

Committee Member

Dampier, David A.

Committee Member

Allen, Edward B.

Date of Degree

5-10-2003

Document Type

Graduate Thesis - Open Access

Major

Computer Science

Degree Name

Master of Science

College

James Worth Bagley College of Engineering

Department

Department of Computer Science

Abstract

A great challenge in the volume-rendering field is to achieve high-quality images in an acceptable amount of time. In the area of volume rendering, there is always a trade-off between speed and quality. Applications where only high-quality images are acceptable often use the ray-casting algorithm, but this method is computationally expensive and typically achieves low frame rates. The work presented here is RZSweep, a new volume-rendering algorithm for uniform rectilinear datasets, that gives high-quality images in a reasonable amount of time. In this algorithm a plane sweeps the vertices of the implicit grid of regular datasets in depth order, projecting all the implicit faces incident on each vertex. This algorithm uses the inherent properties of a rectilinear datasets. RZSweep is an object-order, back-toront, direct volume rendering, face projection algorithm for rectilinear datasets using the cell approach. It is a single processor serial algorithm. The simplicity of the algorithm allows the use of the graphics pipeline for hardware-assisted projection, and also, with minimum modification, a version of the algorithm that is graphics-hardware independent. Lighting, color and various opacity transfer functions are implemented for giving realism to the final resulting images. Finally, an image comparison is done between RZSweep and a 3D texture-based method for volume rendering using standard image metrics like Euclidian and geometric differences.

URI

https://hdl.handle.net/11668/18228

Comments

Scientific visualization||Volume rendering

Share

COinS