Monday, October 3, 2011

Reaction: A Survey of Algorithms for Volume Visualization

This paper is quite an interesting read about volume visualizations which is a fast growing area of scientific visualizations. According to Elvins, typical size of a volume data-set is several hundred megabytes. And to interpret such data on volume visualization is still a challenge for hardware to handle.

Author has clearly explained many of the fundamental concepts of volume visualization along with the other pioneering algorithms. I found this paper quite relevant to understand basics of volume visualizations and to know the generic steps involved in processing volume visualizations.

I particularly find the Ray-Casting algorithm very interesting though I still have few questions about the interpolation that happens to find the value when ray strikes cells between gridpoints. According to the authors too, ray-casting algorithms can be paralyzed at the pixel level since rays from all of the pixels in the image plane can be cast independently. I'm not sure if it is related to the interpolation ?

Animation of volume visualized data can be quite useful for the data analysis process, however in medical field this is critical. Any flaw in the implementation could produce unexpected result and images which may lead to incorrect medical diagnosis. I think with such volume visualization algorithms there's also need to have validation techniques or implementations. Any research on these validations of such algorithms would be very interesting.