Volume visualization of time-varying data using parallel, multiresolution and adaptive-resolution techniques

This paper presents a parallel rendering approach that allows high-quality visualization of large time-varying volume datasets. Multiresolution and adaptive-resolution techniques are also incorporated to improve the efficiency of the rendering. Three basic steps are needed to implement this kind of an application. First we divide the task through decomposition of data. This decomposition can be either temporal or spatial or a mix of both. After data has been divided, each of the data portions is rendered by a separate processor to create sub-images or frames. Finally these sub-images or frames are assembled together into a final image or animation. After developing this application, several experiments were performed to show that this approach indeed saves time when a reasonable number of processors are used. Also, we conclude that the optimal number of processors is dependent on the size of the dataset used