AANS Beyond 2021: Full Collection-Deep Neural networks Can Perform Automated Instrument Detection in Endoscopic Skill Base Surgery

Deep Neural networks Can Perform Automated Instrument Detection in Endoscopic Skill Base Surgery

Video Transcription

Video Summary

In this video, Guillaume Couganer discusses his team's project on automated instrument detection in endoscopic skull-based surgery using deep neural networks. The presentation aims to highlight the use of computer vision in surgical research. He explains the development of an intraoperative video dataset and analytical pipeline, emphasizing the importance of preserving the internal carotid artery in endonasal surgery. The team developed a simulator to train surgeons in managing this complication, recording over 150 trials. They used deep learning models to automatically identify surgical tools in the videos and analyze tool usage patterns in relation to blood loss. The speaker also discusses future applications and challenges in intraoperative video analysis. The dataset used in the project is publicly available for download.

Keywords

automated instrument detection

endoscopic skull-based surgery

deep neural networks

computer vision

surgical research