Feature based opacity function for vessels visualization from CT images

Moti Freiman

Current opacity functions used in volume visualization are adjusted manually according to the iso-values of the visualized structures. However, in CT images of soft tissue it hard to set a good iso-value in order to visualize soft structures like vessels and tumors. Several works pointed out that using multi dimensional opacity function may improve the visualization.
However manual adjustment of a multi dimensional opacity function is very hard. In this work I will present a new algorithm which optimized a multi dimensional opacity function by using the variational principle. More specifically: I will present opacity function which used the geometrical structure of the vessels, and the edges location as input of the opacity function. The geometrical structure computed using multi-scale eigenanalysis of the Hessian matrix. The edges computed using recursive first order Gaussian filters. And finally, The opacity function optimized using the variational approach to minimize energy functions.
Preliminary results will present.

References:
[1] Fattal, R. and Lischinski, D. 2001. Variational classification for visualization of 3D ultrasound data.
   In Proceedings of the Conference on Visualization '01.
   VISUALIZATION. IEEE Computer Society, Washington, DC, 403-410.
   (http://www.cs.huji.ac.il/labs/cglab/research/3dus/Fattal-vis01.pdf).
[2] A. Frangi, W. Niessen, K. Vincken, and M. Viergever. 1998. Multiscale vessel enhancement filtering.
   In W. Wells, A. Colchester, and S. Delp, editors, MICCAI'98, volume 1496 of LNCS, pages 130--137.
   Springer-Verlag, Germany. (http://citeseer.ist.psu.edu/frangi98multiscale.html).