DUDG-SLAM: Dynamic Replay and Depth-Uncertainty-Guided Gaussian SLAM for Robust RGB-D Reconstruction
Jingwen Liu, Tao Zuo, Ai Tian, Yuxin Zhang · 2026
Read DUDG-SLAM: Dynamic Replay and Depth-Uncertainty-Guided Gaussian SLAM for Robust RGB-D Reconstruction by Jingwen Liu, Tao Zuo, Ai Tian, Yuxin Zhang (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13567953.txt.
Description
Jingwen Liu, Tao Zuo, Ai Tian, and Yuxin Zhang present a mapping system that rebuilds indoor spaces from color-plus-depth camera data with fewer errors. Their study was published in twenty twenty-six. Such systems, called SLAM, track the camera while building the scene, and the newer Gaussian-splatting kind can blur as old views fade from memory.
Their method replays past key views chosen by error, forgetting, and visibility, restoring guidance in finished areas. Sensor depth stays the main shape guide, while a learned depth model fills gaps, weighted by how unsure each pixel is. On two standard test sets the system improved image quality, path accuracy, and toughness over its baseline. The work targets robots and headsets that map rooms live.
Catalog record
| CREATOR | Jingwen Liu, Tao Zuo, Ai Tian, Yuxin Zhang |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 64 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |