Scoring in the challenge

141 views
Skip to first unread message

BOCANEGRA PEREZ, ÁLVARO

unread,
Feb 11, 2024, 8:18:34 AMFeb 11
to physionet-...@googlegroups.com
Dear Physionet Challenge Team,

We have some doubts about the objective of the challenge and the way of scoring. When you say that the models must be able to reconstruct the waveforms and/or classify the signals, do you mean that it is possible to perform one of the two tasks or both, or is there some kind of priority (reconstructing the waveforms is mandatory and classification is optional or vice versa), how does this affect the scores?

Thanks,
Álvaro.

PhysioNet Challenge

unread,
Feb 11, 2024, 8:20:33 AMFeb 11
to physionet-challenges
Dear Álvaro,

Yes, there are two separate events/tasks for the 2024 Challenge: ECG digitization and ECG classification. We will compute separate scores for each of these events. Like our previous Challenges, the winners of each of these two events will be the teams whose algorithms achieve the best scores on the hidden test set.

You can participate in the first event, the second event, or both the first and second events. You may find that doing well in the first event helps (or does not help) with the second event, and you may (or may not) want to use parts of your solutions or outputs from the first event for the second event should you participate in both of them. Otherwise, you can consider the events to be independent, and participation in one event is not required for participation in the other event and does not affect the score of the other event.

We have updated the website to distinguish the two events more clearly:
https://physionetchallenges.org/2024/#objective

Best,
Matt
(On behalf of the Challenge team.)

Please post questions and comments in the forum. However, if your question reveals information about your entry, then please email info at physionetchallenge.org. We may post parts of our reply publicly if we feel that all Challengers should benefit from it. We will not answer emails about the Challenge to any other address. This email is maintained by a group. Please do not email us individually.
Reply all
Reply to author
Forward
0 new messages