Compression of Biomedical Images and Signals

Chapter 9

Compression of Static and Dynamic 3D Surface Meshes

9.1. Introduction

Static and dynamic volume data has been used in fluid mechanics, aeronautics, and geology for a long time now, and nowadays it is becoming more widely used in medical imagery to analyze the complex functions such as those in the lungs [PER 04], [FET 03], [FET 05], [SAR 06] and the heart [ROU 05], [ROU 06], [DIS 05] (Figure 9.1).

3D medical imagery is generally visualized by volume rendering [LEV 88]. An observer point of view is selected, its viewing rays cut through the 3D volume perpendicularly to the visualization plane, which is a 2D projection of the volume. The voxels with the same gray level make an isosurface of the same opacity, chosen according to the tissue we wish to view. The gray level gradient allows us to know the normal to the isosurface. Shading based on the transmission, reflection and diffusion of light, gives the desired volume rendering (Figure 9.2a).

An alternative is surface rendering also known as “geometric”. Here, we visualize an isosurface with a predefined gray level. This isosurface is extracted by segmentation of the volume of the data (voxels): a binary volume is constructed. In comparison, volume rendering can display weak surfaces without binary decision. Geometric rendering generally considers the isosurface to be opaque. It reflects the ray according to the normal to the surface, with light diffusion displaying volume rendering. This technique can be used for non-overlapping ...