Track1

2D object detection

For 2D object detection, we provide a real-word training dataset with 10 million unlabled images as well as 5K labeled training set and 5K/10K validation/testing labeled images for evaluation. This dataset has been collected throughout diverse scenarios in more than 30 cities in China and contains scenes in a wide variety of locations, weather conditions and periods, such as highways, city streets, country roads, rainy/snowy weathers, etc.

  • Evaluation: Leaderboard ranking for this track is by Mean Average Precision(mAP) among all categories, that is, the mean over the APs of pedestrian, cyclist, car, truck, tram and tricycle. The IoU overlap threshold for pedestrian, cyclist, tricycle is set to 0.5, and for car, truck, tram is set to 0.7. Only camera images of SODA10M are allowed to be used.
  • Dataset: Please refer to SODA-2d for detailed dataset introduction and dataset downloads.
  • Submission: submit site is now avaliable.

Challenge Results

Below are technique reports of the winner of Track 1:

Track2

3D object detection

For 3D object detection, we provide a large-scale dataset with 1 million point clouds and 7 million images. We annotated 5K, 3K and 8K scenes for training, validation and testing set respectively and leave the other scenes unlabeled. We provide 3D bounding boxes for car, cyclist, pedestrian, truck and bus.

  • Evaluation: Leaderboard ranking for this track is by Mean Average Precision with Heading (mAPH) / L2 among "ALL_NS" (all Object Types except signs), that is, the mean over the APHs of car, cyclist, pedestrian, truck and bus. All sensors are allowed to be used.
  • Dataset: Please refer to ONCE for detailed dataset introduction and dataset downloads.
  • Submission: submit site is now avaliable.

Challenge Results

Below are technique reports of the winner of Track 2:

Track3

Continual Learning:

Supervised classification and 2D object detection
Extension: The deadline for this challenge is extended until October 6th!

For continual learning, we organize 2 supervised subtracks.
Track 3.A focuses on continual classification, providing a natural stream of data from cars driving in China, collected over three days. This challenging track has highly class-imbalanced data (e.g. seeing significantly more cars than pedestrians), and includes domain shifts from day to night in different cities and weather conditions.
Track 3.B focuses on continual 2D Object Detection in a domain-incremental fashion, using the domain shifts in the classification track to group the data into tasks.

  • Evaluation: Leaderboard ranking for track 3A is by Mean Class Accuracy averaged over accross six test points during training. Track 3B is ranked by mAP averaged across all tasks.
  • Dataset: Please refer to SODA-2d for detailed dataset introduction and dataset downloads.
  • Submission: Track 3A and Track B sites are now avaliable.

Challenge Results

Below are technique reports of the winner of Track 3A & 3B:

CHALLENGE PRIZES: (TOTAL 60,000 USD)

Challenge participants with the most successful and innovative entries will be invited to present at this workshop and will receive awards. There are 20,000 USD cash prize for each track. A 10,000 USD cash prize will be awarded to the top performers in each task and 2nd and 3rd places will be awarded with 5 000 USD each. (Subtracks in Track 3 will share the 20,000 USD).


References

[1] Han J, Liang X, Xu H, et al. SODA10M: Towards Large-Scale Object Detection Benchmark for Autonomous Driving[J]. arXiv preprint arXiv:2106.11118, 2021.

[2] Mao J, Niu M, Jiang C, et al. One Million Scenes for Autonomous Driving: ONCE Dataset[J]. arXiv preprint arXiv:2106.11037, 2021.