Progressive approaches of maximum intensity projection and volume rendering for 3D medical images

Abstract

Real-time processing and visualization of 3D image data are the most important features for many applications of medical imaging. Volume visualization is a useful tool to visualize 3D medical images, but large computation amount is a drawback of using a volume-visualized image in medical diagnosis. The processing time of volume visualization depends on the number of voxels of the 3D data. In order to overcome the large amount of computation for the volume visualization, we have developed two progressive visualization methods of a progressive maximum intensity projection and a progressive volume rendering that can perform the volume visualization with low resolution for fast processing and use the low-resolution visualized image to generate a full-resolution visualized image with a reduced computation time. The developed algorithms are applicable to the real-time applications, i.e., the low-resolution visualization is performed interactively for continuous change of view direction, and the full-resolution visualization is performed once the view direction is fixed. Also, the developed algorithms are extensible to multi-level structures. In this dissertation, the progressive visualization methods are implemented and their computation complexities are analyzed. The developed progressive approaches are good solutions for fast processing of the visualization with high image quality.