Implicit Meshes for Effective Silhouette Handling
International Journal of Computer Vision
Understanding 3d emotions through compact anthropometric autoregressive models
ISVC'06 Proceedings of the Second international conference on Advances in Visual Computing - Volume Part I
Graph-based multi-resolution temporal-based face reconstruction
ISVC'06 Proceedings of the Second international conference on Advances in Visual Computing - Volume Part I
| Hi-index | 0.00 |
Deformable 3-D models can be represented either as explicit orimplicit surfaces. Explicit surfaces, such as triangulations orwire-frame models, are widely accepted in the Computer Vision andComputer Graphics communities. However, for automated modelingpurposes, they suffer from the fact that fitting to 2-D and 3-Dimage-data typically involves minimization of the Euclideandistance between observations and their closest facets, which is anon-differentiable distance function. By contrast, implicit surfacerepresentations allow fitting by minimizing an algebraic distancewhere one only needs to evaluate a differentiable field potentialfunction at every data point. However, they have not gained wideacceptance because they are harder to meaningfully deform andrender. To combine the strength of both approaches, we propose amethod that can turn a completely arbitrary triangulated mesh, suchas one taken from the web, into an implicit surface that closelyapproximates its shape and can deform in tandem with it. Thisallows both graphics designers to deform and reshape the implicitsurface by manipulating explicit surfaces using standarddeformation techniques and automated fitting algorithms to takeadvantage of the attractive properties of implicit surfaces. Wedemonstrate the applicability of our technique for upper body-head,neck and shoulders-automated reconstruction.