• All algorithms including deep learning can be applied.
  • Pre-trained models are allowed in the competition.
  • Participants are restricted to train their algorithms on the road damage dataset train sets. Collecting additional data for the target attribute labels is not allowed. If you do so, please mail to the organizers (bdc2018@iis.u-tokyo.ac.jp). Of course, it is permitted to artificially increase training images using data augmentation, GANs and so on.
  • There is no maximum team size
  • You may submit a maximum of 5 submissions per day for judging.
  • Images on this dataset are available under the Creative Commons Attribution-ShareAlike 4.0 International License (CC BY-SA 4.0).Creative Commons License
  • We ask that you respect the spirit of the competition and do not cheat. Hand-labeling is forbidden.
  • Please submit source code and trained model in executable form by November 10. IPython Notebook is desirable for the source code submission, but any programing languages are acceptable. The organizers will verify the reproducibility of the algorithm and determine the final winner.