📢Rules¶

By participating in the COSAS challenge and using the data provided with the challenge, participants agree to adhere to the following participation rules. 

Participation¶

• All participants must form teams (even if the team is composed of a single participant), and each participant can only be a member of a single team. 
• Any individual participating with multiple or duplicate Grand Challenge profiles will be disqualified.
• Anonymous participation is not allowed. To qualify for ranking on the validation/testing leaderboards, true names and affiliations [university, institute (if any) or company (if any), country] must be displayed accurately on verified Grand Challenge profiles, for all participants.
• Members from the organising team and affiliations cannot participate in this challenge. If there is a baseline method in the leaderboard submitted by the organizers, that method will not be taken into account for the final ranking.

Data Use¶

• Every challenge participant agrees to use the provided data only in the scope of CC BY-NC-ND (Attribution-NonCommercial-NoDerivs).
• For training the network, no external data is allowed. Pre-trained networks are permissible only if they have been trained on conventional, non-medical image datasets such as ImageNet or MS COCO.
• Every challenge participant agrees not to make more than one account for downloading data and submitting the results.       

Submission¶

• All final submissions should be in the form of docker containers and follow the template provided by the organizers to be tested successfully.
• It is the user's responsibility to check the sanity of their algorithm's docker during the preliminary test period before submitting it for the final test set.
• Only fully automated methods are acceptable for participation in this challenge as the submission is in the form of docker containers. It is not possible to submit manual annotations or interactive methods.
• Submissions will be ranked solely based on their performance on the final test set and according to the criteria explained in the evaluation page.
• Using pre-trained models on non-medical datasets (like ImageNet, MS COCO) is allowed. Use of external pathology data for model development is not permitted.
• Writing a technical abstract for final submission is compulsory, otherwise the submission will not be considered.
• The deadline for registration and team formation is August 17 2024. Any new team after this deadline will not be considered during the challenge.
• Teams can have 5 submissions on the preliminary test to verify the quality of their methods. However, the number of submissions per day per team and for each task is limited to 1 in order to restrict test data hacking. If a team makes multiple submissions per day (from different users) in the preliminary test phase, they will be disqualified. For clarity, if a team is participating in both tasks, they can make 1 submission per day - one for task 1 and another for task 2. For the final test set, only 1 submission is allowed for each task.