Note: These datasets are intended for educational purposes only. Neither the dataset nor the labels are complete or comprehensive. We encourage you to explore newer, more complete datasets.
The dataset includes driving in Mountain View California and neighboring cities during daylight conditions. It contains over 65,000 labels across 9,423 frames collected from a Point Grey research cameras running at full resolution of 1920x1200 at 2hz. The dataset was annotated by CrowdAI using a combination of machine learning and humans.
- Car
- Truck
- Pedestrian
- xmin
- ymin
- xmax
- ymax
- frame
- label
- preview url for frame
Size | 1.5 GB |
Annotator | CrowdAI |
Caution: The order of the header columns in the labels .csv file in the below Download link is out of order. Please use the corrected .csv file included here in the repository.
This dataset is similar to dataset 1 but contains additional fields for occlusion and an additional label for traffic lights. The dataset was annotated entirely by humans using Autti and is slightly larger with 15,000 frames.
- Car
- Truck
- Pedestrian
- Street Lights
- frame
- xmin
- ymin
- xmax
- ymax
- occluded
- label
- attributes (Only appears on traffic lights)
Size | 3.3 GB |
Annotator | Autti |