Cords: 3D Curve Primitives that Wrap Around Geometry

March 11, 2005
Karan Singh | University of Toronto

The modeling of curves that come in contact with 3D scene geometry presents a challenging problem for users of graphics systems. While embedding curves with physical properties for simulation offers one solution, it does not easily allow an artist or animator to precisely control the final shape, especially when the curves contact complex geometric configurations. This talk addresses the problem with a novel 3D curve primitive called a cord. A cord is defined by a guide curve, which represents an approximate 3D path the user wishes the cord to follow around geometry. The guide curve, along with various attributes, analytically and continuously defines the shape of a cord. The small but intuitive and powerful set of attributes allows users to interactively model and animate curves wrapping around scene geometry. Further, cords can exhibit the visual appearance of a range of materials, including strings, wires, and rubber bands, while allowing animators the precise control of shape needed for keyframe animation. Cords, also find applications in visualization and nonp-hotorealistic rendering and will be shown in the context of the animated short film “Ryan”.

Speaker Details

Karan Singh is an Associate Professor at the University of Toronto, where co-directs the graphics and HCI lab, DGP. His research interests are in artist driven interactive graphics encompassing geometric modeling, character animation and non-photorealistic rendering. He received his BS in Computer Science from IIT, Madras (1991), MS (1992) and PhD (1995) from Ohio State. As a researcher at Alias (1995-1999), he architected facial and character animation tools for Maya. He was involved with conceptual design and reverse engineering software at Paraform (1999-2001) and currently as Chief Scientist for Geometry Systems Inc. He has worked on numerous film and animation projects and most recently was the R+D Director for the Oscar nominated animation Ryan (2005). More details can be found at http://www.dgp.toronto.edu/~karan