{"ID":2863254,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2509.24236","arxiv_id":"2509.24236","title":"PROFusion: Robust and Accurate Dense Reconstruction via Camera Pose Regression and Optimization","abstract":"Real-time dense scene reconstruction during unstable camera motions is crucial for robotics, yet current RGB-D SLAM systems fail when cameras experience large viewpoint changes, fast motions, or sudden shaking. Classical optimization-based methods deliver high accuracy but fail with poor initialization during large motions, while learning-based approaches provide robustness but lack sufficient accuracy for dense reconstruction. We address this challenge through a combination of learning-based initialization with optimization-based refinement. Our method employs a camera pose regression network to predict metric-aware relative poses from consecutive RGB-D frames, which serve as reliable starting points for a randomized optimization algorithm that further aligns depth images with the scene geometry. Extensive experiments demonstrate promising results: our approach outperforms the best competitor on challenging benchmarks, while maintaining comparable accuracy on stable motion sequences. The system operates in real-time, showcasing that combining simple and principled techniques can achieve both robustness for unstable motions and accuracy for dense reconstruction. Code released: https://github.com/siyandong/PROFusion.","short_abstract":"Real-time dense scene reconstruction during unstable camera motions is crucial for robotics, yet current RGB-D SLAM systems fail when cameras experience large viewpoint changes, fast motions, or sudden shaking. Classical optimization-based methods deliver high accuracy but fail with poor initialization during large mot...","url_abs":"https://arxiv.org/abs/2509.24236","url_pdf":"https://arxiv.org/pdf/2509.24236v2","authors":"[\"Siyan Dong\",\"Zijun Wang\",\"Lulu Cai\",\"Yi Ma\",\"Yanchao Yang\"]","published":"2025-09-29T03:20:49Z","proceeding":"cs.RO","tasks":"[\"cs.RO\",\"cs.CV\"]","methods":"[]","has_code":false,"code_links":[{"ID":608990,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_id":2863254,"paper_url":"https://arxiv.org/abs/2509.24236","paper_title":"PROFusion: Robust and Accurate Dense Reconstruction via Camera Pose Regression and Optimization","repo_url":"https://github.com/siyandong/PROFusion","is_official":false,"mentioned_in_paper":false,"mentioned_in_github":true,"github_stars":0}]}