To create a computer vision tool that enables an automatic detection of the division plane around the endometriotic lesion, which requires years of expertise.
A selected few key FEMaLe project outputs and highlights in WP7 from July 2022 to December 2023:
- Legal contracts with 5 health centers world-wide are made.
- 247 confirmed (by experts) surgeries are collected.
- 534 short video sequences are extracted.
- 9,507 endometriosis lesions are totally annotated.
- Annotated zones across 1,320 images with the team of expert and junior surgeons.
- 60 person-hours of discussion to standardize the incision boundary guidelines and reach consensus.
- Annotation pipeline is confirmed based on Delphi method.
- Created a guidebook for data annotations with the effort of surgeons.
- Deep neural network algorithm is designed, implemented and validated.
Noorzadeh et al. Enhancing surgery in endometriosis laparoscopy: training neural networks to segment incision boundaries. Art Int Surg 2024;4:1-6. 10.20517/ais.2024.02.
The algorithm developed marks a substantial leap forward in the realm of surgical technology.
The algorithm will be integrated in the augmented reality (AR) software of SurgAR, a leading innovator in surgical technology.
Integration into SurgAR’s AR platform signifies a transformative advancement in surgical practices, as it empowers surgeons with the ability to identify machine-generated suggestions for incisions in real-time.
This application holds immense promise for enhancing surgical precision and decision-making by providing surgeons with immediate, data-driven insights during procedures.
