AI-Based Real-Time Detection of Surgical Smoke Using Endoscopic Data

Development of AI-Based Approaches for Automated Real-Time Detection of Surgical Smoke Using Endoscopic Image and Video Data

The goal of this observational, prospective monocentric pilot study is to generate a pilot dataset to train a computer-assisted model for automatic, intraoperative detection of surgical smoke gas.

Women with indications for laparoscopic evaluation requiring the use of HF surgery (expecting the formation of smoke gas) and a smoke evacuation system (Karl Storz S-Pilot) are employed.

Detection of surgical smoke gas with an accuracy F1 score of >= 0.8 on test datasets. The activation of the S-Pilot by clinic personnel will be used as the gold standard.

Inclusion Criteria:

• - Age ≥ 18 years
• Written consent after receiving information
• Indication for surgical treatment using HF surgery

Exclusion Criteria:

• - Expected lack of compliance by the patient or inability of the patient to understand the meaning and purpose of the clinical trial
• Lack of patient consent
• S-Pilot cannot be used