SF3D: Stable Fast 3D Mesh Reconstruction with UV-unwrapping and Illumination Disentanglement

Generated objects with SF3D from a single image in under one second.

Abstract

We present SF3D, a novel method for rapid and high-quality textured mesh reconstruction from a single image. Utilizing the Large Reconstruction Model (LRM) as its foundation, SF3D achieves fast generation times with detailed, textured meshes in just 0.5 seconds. Unlike traditional approaches, SF3D is explicitly trained for mesh generation, incorporating a fast UV unwrapping technique that enables swift texture generation rather than relying on vertex colors. The method also learns material parameters and normal maps to enhance the visual quality of the reconstructed models. Furthermore, SF3D integrates a delighting step to effectively remove low-frequency illumination effects, ensuring that the reconstructed meshes can be easily used in novel illumination conditions.

Publication
In ArXiV
Mark Boss
Mark Boss
Research Scientist

I’m a researcher at Stability AI with research interests in the intersection of machine learning and computer graphics.