Challenge schedule

09/13/2023 Start of challenge

Release of training and validation data (with and without ground truth, respectively).

10/13/2023 Release of test data (without ground truth)

Release of test data (without ground truth). At this stage, participants can download an encrypted version of the test-data a few days ahead of the release of the decryption key, to ensure that participants regardless of internet connectivity are on an equal footing.

10/16/2023 Start of test phase (end of development phase)

Release of the decryption key for test data. Participants start predicting the results on the test data. At this stage, they can retrain (or fine-tune) their models while still following the challenge rules.

10/27/2023 End of the Challenge

Deadline for submitting the final predictions over the test (evaluation) data.

10/28/2023 Code and fact sheets submission

  1. Deadline for Code submission with detailed instructions (known as "code verification stage"). This includes a list of requirements, pretrained models, and so on. Note, training code with instructions is also required. Organizers strongly encourage the use of docker to facilitate reproducibility.

  2. In addition to the code, participants are requested to submit the fact sheets using a template provided by the organizers.

Deadline timezone: Anywhere on Earth Time Zone - AOE

Additional details about code and fact sheets submission are provided in our challenge webpage, winning solutions (Post challenge) section.

11/08/2023 Release of final results

11/03/2023 Release of final results

We encourage participants to submit a paper to the associated workshop, independently of their rank position.

  • Paper submission: – October 25th, 2023

  • Paper submission for challenge participants: – November 8th, 2023

  • Decision notification: – November 15th, 2023

  • Camera ready: – November 17th, 2023


WACV'24 Challenge

The ChaLearn WACV'24 Pedestrian Attribute Recognition and Person Retrieval Challenge is open and accepting submissions on Codabench. Join us to push the boundaries of pedestrian attribute recognition along with concept drift.