Identifying different materials within sampled datasets can be an important step in understanding the geometry, anatomy, or

Copyright © 2000 by Academic Press.

All rights of reproduction in any form reserved.

pathology of a subject. By accurately locating different materials, individual parts can be identified and their size and shape measured. The spatial location of materials can also be used to selectively visualize parts of the data, thus better controlling a volume-rendered image [1], a surface model [2], or a volume model created from the data, and making visible otherwise obscured or subtle features. Classification is a key step toward understanding such geometry, as shown in Fig. 1. Figure 2 shows an example of classified MRI data; each color represents a single material identified within the data.

Applications of classified images and geometric models derived from them include surgical planning and assistance, diagnostic medical imaging, conventional computer animation, anatomical studies, and predictive modeling of complex biological shapes and behavior.

0 0

Post a comment