Skullptor: High Fidelity 3D Head Reconstruction in Seconds with Multi-View Normal Prediction

TL;DR

Skullptor introduces a hybrid 3D head reconstruction method combining multi-view normal prediction with inverse rendering, achieving high detail with fewer views and less computation.

Contribution

A novel hybrid approach that extends foundation models with cross-view attention and integrates them into inverse rendering for efficient high-fidelity 3D head reconstruction.

Findings

Outperforms state-of-the-art single-image methods in detail.

Achieves dense-view photogrammetry quality with fewer views.

Reduces computational cost significantly.

Abstract

Reconstructing high-fidelity 3D head geometry from images is critical for a wide range of applications, yet existing methods face fundamental limitations. Traditional photogrammetry achieves exceptional detail but requires extensive camera arrays (25-200+ views), substantial computation, and manual cleanup in challenging areas like facial hair. Recent alternatives present a fundamental trade-off: foundation models enable efficient single-image reconstruction but lack fine geometric detail, while optimization-based methods achieve higher fidelity but require dense views and expensive computation. We bridge this gap with a hybrid approach that combines the strengths of both paradigms. Our method introduces a multi-view surface normal prediction model that extends monocular foundation models with cross-view attention to produce geometrically consistent normals in a feed-forward pass. We then leverage these predictions as strong geometric priors within an inverse rendering optimization framework to recover high-frequency surface details. Our approach outperforms state-of-the-art single-image and multi-view methods, achieving high-fidelity reconstruction on par with dense-view photogrammetry while reducing camera requirements and computational cost.