CAM3R: Camera-Agnostic Model for 3D Reconstruction

TL;DR

CAM3R is a novel camera-agnostic 3D reconstruction model that effectively handles wide-angle and fisheye images without prior calibration, advancing the robustness of 3D scene understanding.

Contribution

It introduces a two-view network with a ray module and cross-view module, along with a ray-aware global alignment for pose refinement, enabling accurate reconstruction across diverse camera models.

Findings

Outperforms existing models on various camera datasets

Achieves state-of-the-art pose estimation accuracy

Successfully reconstructs 3D scenes from wide-angle images

Abstract

Recovering dense 3D geometry from unposed images remains a foundational challenge in computer vision. Current state-of-the-art models are predominantly trained on perspective datasets, which implicitly constrains them to a standard pinhole camera geometry. As a result, these models suffer from significant geometric degradation when applied to wide-angle imagery captured via non-rectilinear optics, such as fisheye or panoramic sensors. To address this, we present CAM3R, a Camera-Agnostic, feed-forward Model for 3D Reconstruction capable of processing images from wide-angle camera models without prior calibration. Our framework consists of a two-view network which is bifurcated into a Ray Module (RM) to estimate per-pixel ray directions and a Cross-view Module (CVM) to infer radial distance with confidence maps, pointmaps, and relative poses. To unify these pairwise predictions into a consistent 3D scene, we introduce a Ray-Aware Global Alignment framework for pose refinement and scale optimization while strictly preserving the predicted local geometry. Extensive experiments on various camera model datasets, including panorama, fisheye and pinhole imagery, demonstrate that CAM3R establishes a new state-of-the-art in pose estimation and reconstruction.