# Online3R: Online Learning for Consistent Sequential Reconstruction Based on Geometry Foundation Model

Source: [Online3R: Online Learning for Consistent Sequential Reconstruction Based on Geometry Foundation Model](<https://www.sota2.com/research/paper/69dc78a57d0172140ec78c29>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-04-14T05:00:11.147Z

Authors: Shunkai Zhou, Zike Yan, Fei Xue, Dong Wu, Yuchen Deng, Hongbin Zha

Published: 2026-04-10T16:42:16.000Z

## Abstract

We present Online3R, a new sequential reconstruction framework that is capable of adapting to new scenes through online learning, effectively resolving inconsistency issues. Specifically, we introduce a set of learnable lightweight visual prompts into a pretrained, frozen geometry foundation model to capture the knowledge of new environments while preserving the fundamental capability of the foundation model for geometry prediction. To solve the problems of missing groundtruth and the requirement of high efficiency when updating these visual prompts at test time, we introduce a local-global self-supervised learning strategy by enforcing the local and global consistency constraints on predictions. The local consistency constraints are conducted on intermediate and previously local fused results, enabling the model to be trained with high-quality pseudo groundtruth signals; the global consistency constraints are operated on sparse keyframes spanning long distances rather than per frame, allowing the model to learn from a consistent prediction over a long trajectory in an efficient way. Our experiments demonstrate that Online3R outperforms previous state-of-the-art methods on various benchmarks. Project page: https://shunkaizhou.github.io/online3r-1.0/

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Camera Pose Estimation on NRGBD](<https://www.sota2.com/research/sota/camera-pose-estimation-on-nrgbd.md>): Camera pose estimation; NRGBD; Error (Breakroom): 0.102.

- [Camera Pose Estimation on TUM RGB-D](<https://www.sota2.com/research/sota/camera-pose-estimation-on-tum-rgb-d.md>): Camera pose estimation; TUM RGB-D; Error (desk): 0.016.

- [3D Reconstruction on 7-Scenes](<https://www.sota2.com/research/sota/3d-reconstruction-on-7-scenes.md>): 3D Reconstruction; 7 Scenes; Completion: 6.7.

- [3D Reconstruction on NRGBD](<https://www.sota2.com/research/sota/3d-reconstruction-on-nrgbd.md>): 3D Reconstruction; NRGBD; Normalized Score (NC): not recorded.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2604.09480>)

[PDF](<https://arxiv.org/pdf/2604.09480.pdf>)
