Haptic Navigation in Volumetric Datasets

Methods of direct or indirect volume rendering are well established and helpful tools for the interactive visual exploration of volumetric datasets. However, the usual display is oriented on surface information, such as the surface of volumetric objects. The actual volumetric information is difficult to interpret intuitively with solely visual methods. In this paper, we add a haptic cue to improve the intuitive interpretation of volumetric data. Based on segmented volume datasets, we generate several distance fields. One distance field encodes the distance from each voxel – which belongs to the segmentation – to the specified surface of the dataset. Another distance field encodes the gradient from a specified starting point to a target point, which is also specified. A local gradient is computed for both distance fields, employing central differences and trilinear interpolation. These gradients are evaluated by a potential field function at the current position. We interpret the resulting vector as a force, which attracts a cursor object to the target position. The resulting drift gives the haptic-enabled user an additional cue to interpret object properties, such as shape, volume, and distances. If a defined vector field is available (i.e., a flow field), low detail vector properties such as large vortices can be ”touched”.