English

LivePose: Online 3D Reconstruction from Monocular Video with Dynamic Camera Poses

Computer Vision and Pattern Recognition 2023-08-22 v2

Abstract

Dense 3D reconstruction from RGB images traditionally assumes static camera pose estimates. This assumption has endured, even as recent works have increasingly focused on real-time methods for mobile devices. However, the assumption of a fixed pose for each image does not hold for online execution: poses from real-time SLAM are dynamic and may be updated following events such as bundle adjustment and loop closure. This has been addressed in the RGB-D setting, by de-integrating past views and re-integrating them with updated poses, but it remains largely untreated in the RGB-only setting. We formalize this problem to define the new task of dense online reconstruction from dynamically-posed images. To support further research, we introduce a dataset called LivePose containing the dynamic poses from a SLAM system running on ScanNet. We select three recent reconstruction systems and apply a framework based on de-integration to adapt each one to the dynamic-pose setting. In addition, we propose a novel, non-linear de-integration module that learns to remove stale scene content. We show that responding to pose updates is critical for high-quality reconstruction, and that our de-integration framework is an effective solution.

Keywords

Cite

@article{arxiv.2304.00054,
  title  = {LivePose: Online 3D Reconstruction from Monocular Video with Dynamic Camera Poses},
  author = {Noah Stier and Baptiste Angles and Liang Yang and Yajie Yan and Alex Colburn and Ming Chuang},
  journal= {arXiv preprint arXiv:2304.00054},
  year   = {2023}
}

Comments

ICCV 2023

R2 v1 2026-06-28T09:43:53.116Z