Poppy provides a training-free way to refine monocular surface normals using single-shot polarization measurements at test time.
March 31, 2026
Original Paper
Poppy: Polarization-based Plug-and-Play Guidance for Enhancing Monocular Normal Estimation
arXiv · 2603.27891
The Takeaway
It allows any frozen RGB backbone (like a Diffusion or Flow-based model) to achieve 25% better accuracy on challenging surfaces (reflective/dark) without retraining. This is a highly practical 'plug-and-play' upgrade for robotics and 3D reconstruction.
From the abstract
Monocular surface normal estimators trained on large-scale RGB-normal data often perform poorly in the edge cases of reflective, textureless, and dark surfaces. Polarization encodes surface orientation independently of texture and albedo, offering a physics-based complement for these cases. Existing polarization methods, however, require multi-view capture or specialized training data, limiting generalization. We introduce Poppy, a training-free framework that refines normals from any frozen RGB