These are some frequently asked questions about the submission:
**Submission form: https://forms.gle/LRtRet96isV1T83r9 **
Q: How do I know the score of my model?
A: you should evaluate your models using the provided data in the Participate > Get Data (https://codalab.lisn.upsaclay.fr/competitions/17705#participate) section in codalab, then submit the best model. The test set consists on that public evaluation data, and other similar clips.
Q: Will the competition be postponed again?
A: No, this is a very difficult competition. Video processing is not supported in the server. The deadline from ECCV to present the material is 31st August. The deadline of the challenge submission form is 25th August.
Q: What is the ranking criteria? VMAF or PSNR?
A: VMAF “uses” (or at least it is correlated with) PSNR-Y. I will ask our collaborators and co-organizers to provide asap evaluation scripts.
Q: Can I submit multiple times?
A: Yes, you can submit through the form multiple times, and edit your submission.
Q: Can I submit to both tracks?
A: of course you can
Q: How are the winners/presentations going to be selected?
A:. The idea is to write a good report and award/give visibility to the most novel solutions, the most efficient solutions, and the top-performing solutions. We depend on our sponsor, Meta, to clarify further this criteria.
We will give at least 2 diplomas (winner and runner-up) based on the three criteria: novelty (subjective, evaluated by video coding experts), efficiency (objective, based on GMACs and/or runtime), and performance (based on VMAF/PSNR-Y).
Best,
Marcos.