Learning Surgical Robotic Manipulation with 3D Spatial Priors
Yu Sheng, Lidian Wang, Xiaomeng Chu, Min Cheng, Bei Hua, Yanyong Zhang, Jiajun Deng, Jianmin Ji · 2026
Read Learning Surgical Robotic Manipulation with 3D Spatial Priors by Yu Sheng, Lidian Wang, Xiaomeng Chu, Min Cheng, Bei Hua, Yanyong Zhang, Jiajun Deng, Jianmin Ji (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/arxiv-2603-03798.txt.
Description
Yu Sheng and seven colleagues give a surgical robot three-dimensional eyes without bolting on extra cameras. Reconstructing the scene first invites error pile-up; wrist cameras, the usual supplement, interfere with the robot's arms and rarely appear in clinics. Their 2026 system, the Spatial Surgical Transformer, reads three-dimensional cues straight from the standard stereo endoscope images, end to end.
The enabler is data: Surgical3D, a photorealistic set of thirty thousand stereo endoscopic pairs with exact geometry, built because surgical three-dimensional data barely existed. A geometric transformer fine-tuned on it extracts robust spatial representations, and a light connector aligns them with the robot's action space in the endoscope's own coordinates. On a real robot the system leads on complex tasks like knot tying and organ dissection outside the body, with strong generalization to new spatial layouts.
Catalog record
| CREATOR | Yu Sheng, Lidian Wang, Xiaomeng Chu, Min Cheng, Bei Hua, Yanyong Zhang, Jiajun Deng, Jianmin Ji |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 36467 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.03798