Texture mapping based Volume Rendering

Ciro Durán <ciro (at) ldc (dot) usb (dot) ve>
Francisco Morillo <francisco (at) ldc (dot) usb (dot) ve>

Abstract
Gallery
Results
Presentations
Bibliography

Abstract

Volume Rendering (VR) is a technique for producing 2D proyections from discreet 3D data. The main applications for this technique are medical or geological data visualization, or fluid dynamics representations. The objective of this thesis is to evaluate the way conventional graphic hardware can be used to produce good quality and high performance visualizations.

VR as a physical process simulates light behaviour passing through a semi-transparent medium, requiring high cost calculations, both in time ans space. New approaches to this problem deal with acceleration the simulation by pre-calculating parts of the process.

We implemented and evaluated several Direct VR techniques, namely ray casting and texture-based renders. We also implemented our own versions of the works of Westermann & Sevenich, and Engel, Kraus & Ertl. The latter produce high quality and interactive framerate images using texture-based algorithms and VR integral pre-calculations (pre-integration), which can compete with more complex techniques.

We obtained a performance and quality reference with the evaluation of direct VR algorithms implemented with OpenGL and Cg (C for Graphics). These results of these implementations were compared against the published results, which show that real-time data visualization with direct VR using conventional graphic hardware is viable.

[Top]

Gallery

**Final screenshots**
Direct VR example	2D Object-Aligned Textures - Skull	3D Object-Aligned Textures - Gray matter	Engel, Kraus & Ertl implementation - Muscle and Skin
Engel, Kraus & Ertl implementation - Fat	Engel, Kraus & Ertl implementation - Skull	Multivolumes - 3 classes: Fat, Skull and Gray matter	Multivolumes - 4 classes: Muscle and Skin, Skull, Fat, Gray matter
Multivolumes - 4 classes: Muscle and Skin, Skull, Fat, Gray matter. Other colors.	Multivolumes - 4 classes: Muscle and Skin, Skull, Fat, Gray matter. Other colors.	Software ray casting - Glial matter	Software ray casting - Gray matter, Muscle and Skin, and White matter.
Westermann after ray casting - Skull	Westermann after ray casting - Muscle and Skin	Differences between eye-aligned and object-aligned slices	Volume and texture proportions
Westermann after ray casting - 2 classes: Gray matter and Fat.	Demonstration of popping between 2 2D Texture sets	Texture overlapping demonstration

**First screenshots**
Gray matter - Full opacity	Gray matter - Half opacity	Gray matter - Other material color	Muscle and Skin - Full opacity
Muscle and Skin - Half opacity	Muscle and Skin - Almost no opacity	Multiple volume software ray casting - Gray matter, Muscle and Skin

**Results Graphics**
Data dispersion	Data occupancy	2D Textures rotating - Zoom 1.0	2d Textures rotating - Zoom 2.0
3D Textures Object-Aligned rotating - Zoom 1.0	3D Textures Object-Aligned rotating - Zoom 2.0	3D Textures Eye-Aligned rotating - Zoom 1.0	3D Textures Eye-Aligned rotating - Zoom 2.0
Multitexture implementation	Engel, Kraus & Ertl implementation	Westermann & Sevenich implementation - Zoom 1.0	Westermann & Sevenich implementation - Zoom 2.0

[Top]

Presentations

We have published some slides about our thesis for an introduction to computer graphics course.

[Top]

Bibliography

Klaus Engel, Martin Kraus, and Thomas Ertl. High-quality pre-integrated volume rendering using hardware-accelerated pixel shading. Eurographics / SIGGRAPH Workshop on Graphics Hardware, 2001. [Site]
Rudiger Westermann and Bernd Sevenich. Accelerated volume ray-casting using texture mapping. IEEE Visualization 2001, 2001.

[Top]

Back to my site

Thesis: Texture Mapping Based Volume Rendering
Authors: Ciro Durán <ciro (at) ldc (dot) usb (dot) ve>, Francisco Morillo <francisco (at) ldc (dot) usb (dot) ve>.
Date: 26 july, 2006, 2:41 A.M.

Texture mapping based Volume Rendering

Contents

Abstract

Gallery

Presentations

Bibliography