Octree based Raycasting for 3D data Visualization

Volumetric data visualization involves the detection of the visible surfaces of an object from a given viewing direction.
The detected surface points are then rendered using known surface visualization techniques. 

We consider the problem of computing the visible surface when the object is given as a three-dimensional array of
intensity values. Traditional methods work directly using the array data structure, or use floating point operations while
using an octree representation. 

In this work, we introduce new ways of using the octree representation and provide evidence to show that an octree
representation is better. Further, our method of discrete raycasting eliminates floating-point computations and is faster
than previous methods.