Papers
38
Total Citations
701
H-Index
17
About
Mobarakol Islam is a pioneering researcher at the intersection of computer vision, deep learning, and robot-assisted surgery, with contributions that are reshaping how intelligent systems perceive and interact with surgical environments. His work spans depth estimation, scene reconstruction, instrument tracking, and surgical language understanding — forming a cohesive body of research aimed at making robotic surgery safer and more intelligent. Islam's most cited work includes **Surgical-DINO** (58 citations), which adapts vision foundation models for depth estimation in endoscopic surgery, and **Endo-4DGS** (45 citations), which pioneers 4D Gaussian Splatting for monocular surgical scene reconstruction. His earlier contributions on spatio-temporal multitask learning and attention-based instrument tracking demonstrated sophisticated approaches to real-time surgical perception under computational constraints. Notably, Islam has also explored the frontier of large language and vision models in surgery, with **SurgicalGPT** and a rigorous empirical study of SAM's applicability in robotic surgery (each garnering 38 citations). His work on surgical report generation further bridges visual understanding with clinical documentation. Collectively accumulating over 400 citations, Islam's research establishes him as a leading voice in AI-driven surgical intelligence, with clear implications for surgical training, augmented reality, and autonomous assistance.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3Learning Where to Look While Tracking Instruments in Robot-Assisted Surgery44 citations · 2019
- 4
- 5
- 6
- 7
- 8
- 9
- 10