Multiple-Volume Volumetric Rendering Using Graphic Accelerator

Abstract

Real-time visualization of volumetric data sets demands high computational performance and large in-memory storage capacity. With the technological advance, the means used to acquire these sets, such as Computed Tomography or Magnetic Resonance Imaging, underwent improvements that resulted in a significant increase of the amount of information collected per sample, making it necessary to develop techniques to generate the visualization of this large quantity of information. The research presents in this paper has the purpose to develop a three-dimensional visualization tool to represent in real time volumetric data with the size of hundreds of megabytes. This tool is based on the ray casting algorithm using graphical acceleration features. To achieve this goal, a methodology was used that involved dividing the volume of data into smaller parts, compatible with the memory limitation of the video hardware. Then the data is used to compose a high quality final image representing the information contained in the volume. In order to evaluate and validate this methodology, comparative tests are performed between the standard rendering procedure and the procedure used in this research.