Hi guys,

by now I have realized that it might be easier to implement my visualization using Tuvok than to investigate what transformations I need to apply to my polygonal surfaces to have them properly aligned with the image in various visualization packages (ImageVis3D svn, 3D Slicer 3.6, MeVisLab 2.1). And if not, implementing will also have a lasting benefit.

Let me first explain my current situation:

I use OpenSceneGraph for visualization, but its volume rendering does not mix properly with its polygonal rendering (if there are overlaps, which is in my case always true), forcing me to resort to displaying isosurface instead of volume rendering. It is ugly and slow. The 3D visualization is inside a widget, and is embedded into Qt main window. Menus and 2D interaction is done using Qt.

I am processing everything in world space in which image is recorded (DICOM defined world space). So positions are usually few hundred units, close to origin (usually the whole image is centered on origin of world space coordinate system), and units are millimeters.

Now how to make the switch in least painful way:

Is there some documentation for Tuvok, some tutorial or something?

If not, can you give me some pointers before I start copying code from iv3d and pruning it. I don't need most of the stuff which in iv3d:

I don't intend to have interface for specifying transfer functions. I think it will be enough to make a few 2D transfer functions with ImageVis3D, save them in files, and then just load them by my program.

If unavoidable, I can convert all my images to UVF format, but I would prefer if I didn't have to do that.

If on the fly change of level of detail is complicated to implement, I can settle for rendering quality switch in the user interface (LQ,MQ,HQ).

Bricking is not very important for me (maybe not needed at all).

A perfect example (if it exists) for me would be qt app with only rendering widget (Tuvok) and 3 buttons: load image, load transfer function, and load triangle mesh.

Thanks,

Dženan