Dynamic harmonic texture mapping using methods of fundamental solutions

Abstract

Feature-aligned surface mapping improves the effect of texture mapping by enforcing feature correspondence between 3D meshes and images. We present a novel constrained texture mapping computation framework based on the method of fundamental solution (MFS). We first flatten the surface onto planar domain, then compose a 2D harmonic map with user specified features aligned with corresponding texels in the texture image. We demonstrate that our framework can efficiently conduct feature-aligned texture mapping with dynamic deforming meshes or texture images.