DriveLM: Driving with Graph Visual Question Answering
Paper
• 2312.14150 • Published
• 5
image imagewidth (px) 627 1.6k |
|---|
This is the sub-dataset of DriveLM which only include single object in ego scenes

For single traffic participant recgonition, segmentation, VQA subtasks of driving scenarios.
single_pedestrian
├── images
└── labeled_pedestrian_data.json
single_vehicle
├── images
└── labeled_vehicle_data.json
Pedestrian
├── token
├── image_path
└── labeled_information
├── gender
├── age_group
└── race
Vehicle
├── token
├── image_path
└── labeled_information
├── type
└── color