High Quality Rendering of Tagged Volumes

Master-Thesis

3D-visualization of tagged volumes on PC-Graphics hardware has been applied within

clinical environments since its introduction. Initially, medical data is explicitly

segmented by experienced physicians. After that, clearly separated areas can be

implicitly segmented through the use of adequate transfer functions for each tag.
The standard visualization approach is based on the use of color tables. Voxel

data contains an index, which is made of the measured values and the corresponding tags.
Transfer functions for each tag are put together in an array and stored as a color

table for texture based visualization. The use of color tables implies pre-interpolative

transfer functions. The hard data boundaries created by the segmentation,

combined with the pre-interpolative transfer functions give a blocky look to the final

volume representation. The aim of this work is to develop an algorithm for interactive

high-quality visualization of tagged volume data in wich the use of post-interpolative

transfer functions is provided and hard data boundaries are smoothed. A possible solution

to this problem can be based on the work from Roettger et al (Vis2003).
(In cooperation with the Neurocenter, Dept. of Neurosurgery)

http://www.nch.med.uni-erlangen.de/forschung/neurozen.php

Advisor(s)

• MSc. CE. Fernando Vega Higuera
• Prof. Dr.-Ing. Stefan Röttger
• PD Dr.-Ing. Peter Hastreiter