Related Experiment Video
Updated: Jul 14, 2026
A Teleoperated Robotic System-Assisted Percutaneous Transiliac-Transsacral Screw Fixation Technique
Published on: January 6, 2023
Development and application of an intelligent platform for automated recognition of surgical instruments in
Ran Hu1,2, Ming Tang1,2, Yi Jin1,2
1Department of General Surgery, Division of Pancreatic Surgery, West China Hospital of Sichuan University, Chengdu, China.
Background:
Surgical videos offer rich intraoperative data critical for surgical education, quality control, and skill assessment. However, their increasing volume and complexity render manual review impractical. To address this challenge, this study aims to develop an intelligent platform capable of automatically recognizing and visualizing intraoperative surgical instrument usage, and to evaluate its clinical utility in real-world surgical settings.
Materials And Methods:
Surgical videos from 21 medical centers in China, covering five surgical types, were collected to develop a generalized artificial intelligence (AI) model for automated surgical instrument recognition. The model was deployed on the SurgSmart platform which features clinically oriented functions. A multicenter survey involving 30 surgeons was conducted to assess the clinical value of the platform.
Results:
A total of 1261 surgical videos were collected, from which 96 324 images were extracted and annotated with 268 828 labels. The developed model achieved a mean Average Precision of 80.31% for recognizing 21 surgical instruments. Based on this model, four core functions were implemented on SurgSmart: Rapid Review Mode, Surgical Instrument Report, Surgical Instrument Heatmap, and Surgical Teaching Mode. All participating surgeons reported a high level of satisfaction and acknowledged the clinical relevance of these functionalities.
Conclusion:
A universally applicable surgical instrument recognition model was developed and deployed on SurgSmart to enable the visualization of intraoperative instrument usage, demonstrating promising clinical potential for automated surgical video analysis and enhanced intraoperative data interpretation.
