Visualization: A Signal Processing Approach to Direct Volume Rendering

In this project we explore the application of state-of-the-art signal processing techniques into volumetric visualization to extract additional information to provide more knowledge about the content inside the dataset.

One approach we have investigated is use shape priors and adaptive filters to automatically adjust the Transfer Functions according to variations in the data, such as varying concentrations of contrast agent in CT Angiography [1]. Another approach we have investigated is to include knowledge from the user already in the reconstruction step of the volume rendering pilpeline. By doing so, we are able to prevent artifacts in the form of of falsely classified samples due to intepolation effects that arise from the assumption of data continuity.

[1] G. Läthén, S. Lindholm, R. Lenz, M. Borga. Automatic Tuning of Spatially Varying Transfer Functions for Blood Vessel Visualization. In IEEE Transactions on Visualization and Computer Graphics, 12(18):2345-2354, 2012.