Papers
21
Total Citations
261
H-Index
10
About
Mengya Xu is an emerging researcher at the forefront of computer-assisted intervention and surgical AI, with expertise spanning scene reconstruction, semantic segmentation, surgical report generation, and robotic surgery understanding. Her work addresses critical challenges in making robot-assisted surgery safer, smarter, and more interpretable through cutting-edge deep learning techniques. Among her most influential contributions is "Endo-4DGS" (45 citations), which pioneers 4D Gaussian Splatting for dynamic endoscopic scene reconstruction — a significant leap forward in surgical scene modeling. Her empirical investigations into the Segment Anything Model (SAM) within robotic surgery contexts have garnered considerable attention (38 and 12 citations respectively), rigorously evaluating foundation model generalizability in clinical domains. Her paired works on surgical report generation with domain adaptation and calibration (29 and 27 citations) demonstrate her sustained focus on bridging surgical scene understanding with natural language outputs, directly supporting documentation and training workflows. Xu has also advanced surgical instrument segmentation, fine-grained interaction recognition, and privacy-preserving continual learning for surgical semantics. With over 220 cumulative citations across diverse problem spaces, her research consistently tackles real-world deployment challenges in minimally invasive and robotic surgery, making her a compelling voice in surgical AI innovation.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5
- 6SIRNet: Fine-Grained Surgical Interaction Recognition16 citations · 2022
- 7
- 8
- 9SAM Meets Robotic Surgery: An Empirical Study in Robustness Perspective12 citations · 2023
- 10