Sparse voxel representation of point cloud data

Is there a c/c++ implementation of sparse voxel representation开发者_JS百科 for any 3D point cloud dataset?

Voxlap by Ken Silverman is perhaps the closest you can get.

I imagine that the medical imagers use proprietary engines. I'd further imagine that the model is in fact the domain model, and the renderer isn't in visiting voxels but rather visiting cells which it can derive a colour value from. So its not some generic piece of code that gets shared.

And other voxel engines are mostly for landscape, and increasingly in Flash. I saw a neat one recently here: http://planeman-bluffersguide.blogspot.com/2005/01/testing-swf.html

C&C Tiberian Sun used voxel models, and encoded 'skips' in one dimension so the renderer was never checking to see if cells were empty, but rather just visiting the busy cells.

With the interest in voxels by ID Software, octtree and other space dividing algorithms are being considered. RTree would also be useful, perhaps.