Fast segmentation, conversion and rendering of volumetric data using GPU

In this paper we present a proof-of-concept implementation of fast volumetric data segmentation, conversion to polygonal mesh geometry and rendering. All parts of the method are implemented on the graphical processing unit, which allows high degree of parallelization. Implementations of presented algorithms are done in the OpenCL framework and are integrated in blood vessel visualisation software Neck Veins. This paper presents where and to what degree parts of method can be parallelized. In results we also show to what degree we can speed-up the implementation by using parallel computing power of the graphical processing units.