Prof. D. Silver, Dept. of Electrical and Computer Engineering, Rutgers, The State University of New Jersey
Research Abstract:
3D data (from 3D imaging, simulations, sensing) is prevalent in many scientific fields such as medicine, biology, engineering, psychology, security & screening, and archeology. However, the effective exploration of 3D data is still a challenging task. Especially difficult is the planning of a complex 3D procedure based solely on the limited interaction available, such as rotation and slicing. In science, medicine and engineering, hand-drawn illustrations often include manipulating part of an object to depict the stages and the outcomes of a procedure or process. These illustrations help to uncover hidden features or reveal the spatial relationship between different components of the object. This research involves the exploration of a more "active" set of visualization techniques for 3D that mimic styles in illustrations and which will allow querying of 3D data using manipulations such as cutting, pulling and peeling.
The techniques are based on the concept of "illustrative deformation". Illustrative deformation is a non-physically-based means for specifying and realizing visualizations that contain cuts, dissections, distortions, and various other forms of deformation that are not present in the original data. The deformation is represented through iconic operators which are applied to a 3D dataset much like a texture is applied to a polygonal model. By using inverse rendering and correct normal calculation of the newly created surfaces, high-quality illustration-like images can be achieved at interactive speeds. This research is investigating, developing and validating a 3D interaction paradigm through illustrative deformation.
