Generic deformable implicit mesh models for automated reconstruction

Abstract

Deformable 3-D models can be represented either as explicit or implicit surfaces. Explicit surfaces, such as triangulations or wire-frame models, are widely accepted in the computer vision and computer graphics communities. However, for automated modeling purposes, they suffer from the fact that fitting to 2-D and 3-D image-data typically involves minimization of the Euclidean distance between observations and their closest facets, which is a non-differentiable distance function. By contrast, implicit surface representations allow fitting by minimizing an algebraic distance where one only needs to evaluate a differentiable field potential function at every data point. However, they have not gained wide acceptance because they are harder to meaningfully deform and render. To combine the strength of both approaches, we propose a method that can turn a completely arbitrary triangulated mesh, such as one taken from the Web, into an implicit surface that closely approximates its shape and can deform in tandem with it. This allows both graphics designers to deform and reshape the implicit surface by manipulating explicit surfaces using standard deformation techniques and automated fitting algorithms to take advantage of the attractive properties of implicit surfaces. We demonstrate the applicability of our technique for upper body-head, neck and shoulders-automated reconstruction.