For this challenge your team will use videos from robotic surgical training exercises along with labels of the tools present and surgical tasks being performed, to develop algorithms for surgical tool detection and task recognition. A large data set of 280 long surgical videos (comprising of over 18 million images) along with corresponding labels is provided.
In order to participate in this challenge, please follow the steps below:
- Register for the challenge by clicking on the "Join" button on top right of this webpage
- Read and agree to the challenge guidelines here (which are expanded version of the agreement form from step 2).
- The team lead (only) should send the agreement form available here to isi.challenges@intusurg.com after having all team members request to join the challenge as well.
- Once your team members are accepted into the challenge, create your team in the challenge.
- Familiarize yourself with the timeline for the challenge here.
- Read and understand the two challenge categories here.
- Understand the details of training and testing data here.
- Check out the prizes for the challenge here.
- Download data from here. (only accessible after completing registration steps)
- Understand the evaluation criteria as described here.
- Make sure you understand the submission instructions available here.
- Review SurgToolLoc 2022/23 resources here for context and to understand what worked for winning teams and what didn't (in regards to category 1 of this challenge)
- Optional: Check out this great tutorial developed by the NVIDIA team from SurgToolLoc 2022 challenge.
- Make your submissions directly online to the grand-challenge website along with sending necessary submission material (methodology report, github repo, pre-recorded method presentation)
- Register for MICCAI and attend EndoVis!