Driving conditions in India are highly unstructured and diverse, with interesting behaviors of traffic participants, compared to the rest of the world. These driving conditions pose unique challenges that are yet unsolved, for research in artificial intelligence (AI) and machine learning (ML) systems, and hence offer immense opportunities for possible technical innovations in AI/ML systems for better road safety.
India Driving Dataset (IDD) is the world’s first dataset of unstructured driving scenarios. This is an open dataset, available for anyone across the world, for research. IDD was released at the first AutoNUE workshop in the European Conference on Computer Vision (ECCV) 2018 with 50,000 images of ground truth annotation. In the second AutoNUE workshop at the International Conference on Computer Vision ( ICCV) 2019, IDD is expanded to be a multi-modal dataset with more data from stereo camera, LIDAR, and GPS from diverse road conditions.
As part of the NCVPRIPG 2019, we are launching a data challenge:
This challenge is open to students in Indian universities and colleges. The challenge will use IDD-Lite, a dataset that is small and compact to fit on any personal computer and so will not huge-compute infrastructure. IDD-Lite is less than 50MB is size, contains 7 classes (compared to 30 in IDD). The dataset and the evaluation code is already available for participants. An automated leaderboard will be made available on 25-November-2019. We will also be releasing a sample code which gives a baseline accuracy, for the challenge.
Intel is sponsoring the prizes to students - cash prizes to leading entries (or teams), and non-cash prizes as travel grants or conference registration to limited number of select participants in the data challenge.
python evaluate/idd_lite_evaluate_mIoU.py --gts $GT --preds $PRED --num-workers $C