Skip to yearly menu bar Skip to main content


Rethinking Inductive Biases for Surface Normal Estimation

Gwangbin Bae · Andrew J. Davison

Arch 4A-E Poster #95
[ ]
Thu 20 Jun 10:30 a.m. PDT — noon PDT
Oral presentation: Orals 3A 3D from single view
Thu 20 Jun 9 a.m. PDT — 10:30 a.m. PDT


Despite the growing demand for accurate surface normal estimation models, existing methods use general-purpose dense prediction models, adopting the same inductive biases as other tasks. In this paper, we discuss the inductive biases needed for surface normal estimation and propose to (1) utilize the per-pixel ray direction and (2) encode the relationship between neighboring surface normals by learning their relative rotation. The proposed method can generate crisp - yet, piecewise smooth - predictions for challenging in-the-wild images of arbitrary resolution and aspect ratio. Compared to a recent ViT-based state-of-the-art model, our method shows a stronger generalization ability, despite being trained on an orders of magnitude smaller dataset.

Live content is unavailable. Log in and register to view live content