Next: References Up: Title
Page Previous: Browsing 3D Models
Index: Full Text Index Contents: Conference
Page
Conclusions
Thanks to the foresight of the National Library
of Medicine, there exists one of the largest, most coherent collection
of anatomical image data anywhere, and we can be challenged to test our
visualization tools against the data. Simply manipulating the data is a
challenge. In addition, the data is highly complex and we desperately need
effective and efficient methods of extracting the visible structures. Therefore
we should focus on developing new segmentation algorithms which could separate
most soft-tissue structures from the surrounding or adjacent tissues.
The Level Of Detail (LOD) present in the original
Visible Human data should be preserved in the visualization process and
passed to the 3D models. We believe that our algorithm where we build
very detailed surface-based models which are "dipped" into their corresponding
volume of interest to get the color for the vertices, captures precisely
both the shape and the texture of the structures. We believe that it is
feasible to generate highly photographic quality 3D models of anatomical
structures which can be tailored to low end computer platforms. The multi-resolution
mesh representation with perceptually preserved textures is one way of
approaching the problem.
Next: References Up: Title
Page Previous: Browsing 3D Models
Index: Full Text Index Contents: Conference
Page
