Dataset Features

Accident Scenarios

Diverse accident scenarios across different places under various weather and lighting conditions.

LiDAR
Accident moment
LiDAR
Accident moment
LiDAR
Accident moment

Multi-Agent Data

Synchronized data between four V2X vehicles and one V2X infrastructure to provide diverse viewpoints.

LiDAR
Collaborative autonomous driving
LiDAR
V2X settings of DeepAccident

Multi-Modality Sensors

Multi-modality sensors including LiDAR and multi-view RGB cameras on each V2X agent.

LiDAR
LiDAR
Camera
Camera

Diverse Annotations

Ground truth labels for motion and accident prediction, semantic segmentation, and 3D object detection and tracking

LiDAR
Motion & accident prediction
LiDAR
BEV semantic segmentation
LiDAR
Bounding box on LiDAR
LiDAR
Bounding box on image

Diverse Scenarios


Accident Scenario Visualization

BEV Semantic Segmentation


Multi-View Cameras



Motion and Accident Prediction Task

V2X Settings

A description of your image
  • For each scenario, we have five V2X agents:
    • Two vehicles designed to collide with each other
    • Two vehicles followed behind the colliding vehicles.
    • One infrastructure equipped with full set of sensors
  • V2X Settings

    (for illustration, the cameras are installed behind vehicles)

    Vehicle Side

    Instructure Side


    Multi-Modality Sensors

    A description of your image

    Citation

    If you find the dataset helpful, please consider citing us.

    @article{Wang_2023_DeepAccident,
        title = {DeepAccident: A Motion and Accident Prediction Benchmark for V2X Autonomous Driving},
        author = {Wang, Tianqi and Kim, Sukmin and Ji, Wenxuan and Xie, Enze and Ge, Chongjian and Chen, Junsong and Li, Zhenguo and Ping, Luo},
        journal = {arXiv preprint arXiv:2304.01168},
        year = {2023}
    }