MAFAT Radar Challenge has recently ended. This was the second competition in the MAFAT Challenge series. The competition focused on the classification of living, non-rigid objects, detected by doppler-pulse radar systems.
We've had more than 1,000 registered participants and more than 4,300 submissions. During the competition, participants have received a training data set containing 6,656 radar segments, labeled as either animals or humans, in addition to a supportive (auxiliary) data set containing 49,071 segments. Participants trained their machine learning predictive models based on the training data, and were asked to make predictions for this binary classification task on untagged data (the test set).
After the private test phase has come to its conclusion, the participants' submissions were evaluated against the true labels of the private test set (labels that were unknown to the participants), yielding the following results:
Since the competition ended, we have been evaluating the submissions of the leaders and verifying their prize-winning eligibility.
1st place ($25,000) prize winner – GSI Technology
Today, we are delighted to formally announce GSI Technology as the competition winners!! They have put a lot of effort and creativity into this challenge, and they have passed all our formal and technical eligibility tests.
If you want to study their technical approach in this challenge, take a look at this blog post that was published by Daphna Idelson from GSI.
2nd place ($10,000) prize winner – Axon Pulse
We would like to congratulate Axon Pulse on a second-place finish, their technical approach to this challenge is described in the following blog post.
3rd place ($5,000) prize winner – Ido Kazma
Some of you might recognise his Codalab leaderboard username “rogueneuron”, congratulations Ido!
MAFAT would like to thank all of the participants for embracing the challenge and for their enthusiastic participation and cooperation!
We would like to remind you that MAFAT plans to launch additional competitions. You can subscribe to MAFAT Challenge's mailing list here and get information regarding future competitions straight to your inbox.