The quality of a 3D volume visualization heavily depends on a representative transfer function which is responsible for mapping the original density values to color and opacity. Finding a suitable transfer function is often a tedious task if done manually in a trial-and-error fashion by specifying piecewise linear functions for each color and opacity channel. Contrary, image-based models exploring features like gradient magnitude, histogram, or edge detection do not consider much user interaction as performed almost autonomously. Hence, we propose a new paradigm which integrates the user into the transfer function specification process. This allows an intuitive specification within an Augmented Reality environment by providing different predefined shape functions which can easily be adjusted. Moreover, a new approach is introduced which utilizes spatial information as an additional transfer function. This opens a completely new way of exploration in volume visualization. Keywords tran...