Master Thesis seminar: Generation of smooth non-manifold surfaces from segmented image data

Date: May 19, 2004 (Wednesday) at 15:15

Av: Jon Adolfsson och Johan Helgesson
Projektet är utfört på Konrad Zuse-institutet i Berlin

This master's thesis treats the problem of correctly visualizing scientific data sets that contain many different materials. The studied data sets come from the medical applications where many different tissues within a patient have to be visualized simultaneously. These data sets are often acquired when using MR- or CT-scanner, devices more and more common as standard hospital equipment.

A visualization system called Amira has been developed at ZIB in Berlin to visualize these kinds of medical data. Amira visualize the tissues by generating three-dimensional surface meshes of triangles. Many different tissues can be visualized at the same time. Where three or more tissues meet
some ambiguities arise of how to visualize the connections between the tissues correctly. In the current version of Amira all materials are treated in the same way which unfortunately leads to ridges and malformations on the surface when hard and smooth materials like bone are visualized.

The master's thesis deals with two different approaches to remove these ridges from the surface. The first approach modifies the way the surface construction algorithm in Amira works in order to directly generate one of the materials as smooth. The second approach deals with the output surface from the surface construction algorithm and removes the ridges by identifying their position and applying a low-pass filter.

Room: E:1147

Last modified Dec 9, 2011 12:59 pm