We are excited to announce the CholecTriplet2022 challenge, a surgical action triplet detection challenge which is part of the Endoscopic Vision Grand challenge at MICCAI 2022. In this edition, participants are tasked with recognizing and localizing tool-tissue interactions, represented as triplets of <instrument, verb, target>, from laparoscopic videos. Participants will compete to train models using only frame-level annotations to predict the triplets present in a given image and localize the instrument used to carry out the action. 

This challenge is in line with recent efforts in the computer vision and deep learning community, which have begun to transition from coarse-grained tasks such as action and activity recognition to finer-grained tasks such as Human-Object Interaction (HOI).  Formalizing surgical activities using triplets and localizing them represents a step toward comprehensive fine-grained modeling in the surgical domain, and can facilitate the development of intra-operative decision support systems in the operating room (OR). CholecTriplet2022 offers participants the opportunity to explore these research areas, engage with the community, compete, collaborate, and push the boundaries of surgical scene understanding.

This challenge will provide access to the CholecT50 dataset, a dataset of 50 surgical videos which has been annotated by clinicians with 100 action triplet classes. A baseline study [1] provides additional background about the task and dataset. The teams with the best results will be rewarded with prizes. We also plan a joint publication with top participants after the challenge.

To register for the challenge and for more information, please visit the challenge website: 

https://cholectriplet2022.grand-challenge.org/

Please feel free to contact us at the email below should you have any questions. We look forward to your participation!

Best regards,

The CholecTriplet2022 Organizers.

Aditya Murali, Chinedu Nwoye, Saurav Sharma, Tong Yu, Armine Vardazaryan, Deepak Alapatt, Nicolas Padoy

✉ Contact us: cholectriplet2022-support@icube.unistra.fr

[1]Nwoye, C. I., Yu, T., Gonzalez, C., Seeliger, B., Mascagni, P., Mutter, D., … & Padoy, N. (2022). Rendezvous: attention mechanisms for the recognition of surgical action triplets in endoscopic videos. Medical Image Analysis, 102433.