SLAM3R: Real-Time Dense Scene Reconstruction from Monocular RGB Videos
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866917965927022592 |
|---|---|
| author | Liu, Yuzheng Dong, Siyan Wang, Shuzhe Yin, Yingda Yang, Yanchao Fan, Qingnan Chen, Baoquan |
| author_facet | Liu, Yuzheng Dong, Siyan Wang, Shuzhe Yin, Yingda Yang, Yanchao Fan, Qingnan Chen, Baoquan |
| contents | In this paper, we introduce SLAM3R, a novel and effective system for real-time, high-quality, dense 3D reconstruction using RGB videos. SLAM3R provides an end-to-end solution by seamlessly integrating local 3D reconstruction and global coordinate registration through feed-forward neural networks. Given an input video, the system first converts it into overlapping clips using a sliding window mechanism. Unlike traditional pose optimization-based methods, SLAM3R directly regresses 3D pointmaps from RGB images in each window and progressively aligns and deforms these local pointmaps to create a globally consistent scene reconstruction - all without explicitly solving any camera parameters. Experiments across datasets consistently show that SLAM3R achieves state-of-the-art reconstruction accuracy and completeness while maintaining real-time performance at 20+ FPS. Code available at: https://github.com/PKU-VCL-3DV/SLAM3R. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2412_09401 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | SLAM3R: Real-Time Dense Scene Reconstruction from Monocular RGB Videos Liu, Yuzheng Dong, Siyan Wang, Shuzhe Yin, Yingda Yang, Yanchao Fan, Qingnan Chen, Baoquan Computer Vision and Pattern Recognition In this paper, we introduce SLAM3R, a novel and effective system for real-time, high-quality, dense 3D reconstruction using RGB videos. SLAM3R provides an end-to-end solution by seamlessly integrating local 3D reconstruction and global coordinate registration through feed-forward neural networks. Given an input video, the system first converts it into overlapping clips using a sliding window mechanism. Unlike traditional pose optimization-based methods, SLAM3R directly regresses 3D pointmaps from RGB images in each window and progressively aligns and deforms these local pointmaps to create a globally consistent scene reconstruction - all without explicitly solving any camera parameters. Experiments across datasets consistently show that SLAM3R achieves state-of-the-art reconstruction accuracy and completeness while maintaining real-time performance at 20+ FPS. Code available at: https://github.com/PKU-VCL-3DV/SLAM3R. |
| title | SLAM3R: Real-Time Dense Scene Reconstruction from Monocular RGB Videos |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2412.09401 |