MAFAT Fine-Grained Classification Challenge – Private Leaderboard

The first competition in the MAFAT Challenge series has recently ended. The competition focused on fine grained classification of objects from aerial imagery.

During the competition, participants have received a training data set containing 1,663 high resolution images, marked with 11,617 vehicles. Each vehicle in the training data set was labeled with fine grained details, including its class, sub class, unique features and perceived color. Participants trained their machine learning predictive models based on the training data, and were asked to make predictions of those fine grained details on untagged data (the test set).

It's time to publish the private leaderboard. The final submissions of individuals / teams were graded on the private test set, yielding the following results:

Congratulations to the leaders!

We are still evaluating the leaders' submissions and verifying their prize winning eligibility. Once this process is done, the leaders' status will be officially changed to WINNERS… Then, we will publish interviews with the winners.

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 in 2019, so stay tuned to for details.