An evaluation of the Volume on Surface (VoS) approach

Abstract

Hybrid volume rendering algorithms combine techniques of different categories of volume visualization. Volume on Surface (VoS) is a hybrid volume rendering technique that maps volume information to isosurfaces, intending to accelerate the volume rendering process. In this paper, we introduce an improved version of the VoS technique. Furthermore, we conduct an evaluation of CUDA-based versions of both the original and the improved approach, comparing them with the conventional Ray Casting algorithm. In some results, our technique is more than eight times faster than Ray Casting and thirty times than original version. In addition to the main discussion, we investigate how a surface simplification influences the performance and the quality of the images rendered by our approach.