Visualizing uncertainty in multi-resolution volumetric data using marching cubes

Data sets acquired from complex scientific simulation, high precision engineering experiment and high-speed computer network have been exponentially increased, and visualization and analysis of such large-scale of data sets have been identified as a significant challenge to the visualization community. Over the past years many scientists have made attempt to address this problem by proposing various data reduction techniques. Consequently the size of data can be reduced and issues associated to the visualization can be improved (e.g. real-time interaction and visual overload). However, during the process of data reduction, the information of original data sets was approximated and potential errors were introduced. It leads to a new problem with regard to the integrity of the data and might mislead users for incorrect decision making. Therefore in this paper we aim to solve the problem by introducing three novel uncertainty visualization methods, which depict both the multi-resolution (MR) approximations of the original data set and the errors associated with each of its low resolution representations. As a result we faithfully represent the MR data sets and allow users to make suitable decisions from the visual output. We applied our techniques on a data set from medical domain to demonstrate their effectiveness and usability.