This thesis combines a 3D interaction model with a Maximum Intensity Difference Accumulation (MIDA) volume visualization algorithm to create a technique for exploring volumetric datasets. The interaction model is based on a 3D Painting metaphor where a user selects a Region of Interest (ROI) by “painting” a 3D envelope enclosing features of interest. The result is an exploration technique that is intuitive to use and easy to learn even for non-expert users. The painting based model and the MIDA algorithm also provide visualization flexibility by allowing for different combinations of volumetric exploration operations. In addition, the various algorithms comprising the exploration technique have been implemented to take full advantage of parallel computational capabilities of modern Graphics Processing Units (GPUs), thus providing real-time interaction and high-quality visualisation. Finally, the contributions of the thesis are validated by a series of experiments and a user study.