Redefining the Laparoscopic Spatial Sense: AI-based Intra- and Postoperative Measurement from Stereoimages

• DOI: 10.48550/arxiv.2311.09744
• Publication Date: 2023-01-01
• Authors: Leopold Müller; Patrick Hemmer; Moritz Queisner; Igor M. Sauer; Simeon Allmendinger; Johannes Jakubik; Michael Vössing; Niklas Kühl
• Keywords: Robustness

A significant challenge in image-guided surgery is the accurate measurement task of relevant structures such as vessel segments, resection margins, or bowel lengths. While this an essential component many surgeries, it involves substantial human effort and prone to inaccuracies. In paper, we develop a novel human-AI-based method for laparoscopic measurements utilizing stereo vision that has been guided by practicing surgeons. Based on holistic qualitative requirements analysis, work proposes comprehensive method, which comprises state-of-the-art machine learning architectures, RAFT-Stereo YOLOv8. The developed assessed various realistic experimental evaluation environments. Our results outline potential our achieving high accuracies distance with errors below 1 mm. Furthermore, on-surface demonstrate robustness when applied challenging environments textureless regions. Overall, addressing inherent challenges surgery, lay foundation more robust solution intra- postoperative measurements, enabling precise, safe, efficient surgical procedures.