Hellas,
the visualiztion technique is called 'Volume Rendering', the wikipedia article has quite some information on it:
Also talks about optimizations etc. The only two techniques of which I know how they work are
1. slices based approach, where you render a stack of semi-transparent textures oriented towards the user, slicing through the volume data and mapping the scalar field value to texture color with what is called a lookup-table (LUT)
2. raycasting based approach, where you sample the volume data along sight rays originating in your camera center and going through each pixel on your projection place. This approach I think has more potential for optimization since you might terminate your rays early, can be better parallelized using fragment shaders for the per-pixel compuation, allows e.g. supersampling/antialiasing and can be dynamically adapted for a quality / performance tradeoff
You might conduct additional research about more recent state of the art.
There is an implementation in VTK afaik (http://vtk.org).
Edit: Oh yeah, this one: http://www.amazon.com/Real-Time-Volume-Graphics-Klaus-Engel/dp/1568812663
Have fun!
Cheers
Patric
