The submission should contain a technical report containing the description of the method and .txt files containing the results for the testing data. The technical report should not exceed 10 pages of MICCAI paper format. Please read carefully the terms and conditions of submission at the end of this page.

The name of the result files should be as follows: <VideoName>_result.txt. All the annotation files are generated at 1 fps, thus the result files should also be generated at 1 fps. The content of the result files should follow this convention:

Frame Grasper Bipolar Hook Scissors Clipper Irrigator SpecimenBag
<frame ID  0>  <confidence vector>
<frame ID 25>  <confidence vector>
<frame ID 50>  <confidence vector>
<frame ID 75>  <confidence vector>


Frame Grasper Bipolar Hook Scissors Clipper Irrigator SpecimenBag
0 -1.5348 -3.0689 2.8234 -0.9753 -3.3533 -2.8772 -1.2033
25 2.8949 1.0443 0.38031 -3.6901 4.5647 -0.74929 -3.1698
50 -2.6808 1.3377 0.7014 -2.6383 0.20482 -4.2711 0.21033
75 -1.6789 0.20425 1.3347 1.3916 2.6992 2.7968 -3.2456

Please compress your results and the report into a single zip archive with the naming format: <LastName>_<Institution> To obtain the data, please fill the form.

The submission for the challenge is not anonymous, so it is allowed to put your names and affiliations in the report. Any supplementary materials can be included in the .zip file inside a folder named “supplementary”. The reception of submission will be confirmed by an email response.

Further details regarding the submission procedure will be provided later.

Terms and Conditions of Submission

By submitting their results, the participants understand and agree on these terms and conditions:

  • At least one of the participants from each submission commits to attending the M2CAI event.
  • We are going to maintain a list which ranks the performance of the methods involved in this challenge. The technical reports and the results of the 3-best performing methods will be published on the website. The authors from all top-3 submissions will also give an oral presentation at the M2CAI event.
  • Other participants have the choice whether or not to publish their method and results on the website. Depending on the venue, they might also be asked to bring a poster of their proposed approach to the M2CAI event.
  • The submitted report contains a clear description of the methodology, including (but not limited to) the parameters required by the approach and additional external data used during training (e.g., extra videos from other dataset, additional labels added to the training data). 
  • The test dataset has not been used for training the method in any way.
  • The content of the result file is generated using the method described in the report and is not modified through any means of manipulation.
  • [Optional] We strongly encourage the participants to also write their report in a full paper format (including abstract, introduction, related work, etc.) and publish it on ArXiv, to facilitate comparisons in future work.