Skip to content

Latest commit

 

History

History
47 lines (39 loc) · 6.43 KB

dataset_summary.md

File metadata and controls

47 lines (39 loc) · 6.43 KB
Raw

A summary of large-scale collaborative perception datasets

Usage of Common Datasets (from papers with code):

Dataset Venue Source Frame V2V V2I I2I Agents Camera LiDAR Depth OD SS OT MP Website
V2V-Sim [1] ECCV'20 Simu 51K ✔ - - 1-7 - ✔ - ✔ - - ✔ -
V2X-Sim [2] RAL'21 Simu 10k ✔ ✔ - 1-5 ✔ ✔ ✔ ✔ ✔ ✔ - Link
OPV2V [3] ICRA'22 Simu 11K ✔ - - 1-7 ✔ ✔ - ✔ ✔ - - Link
DAIR-V2X-C [4] CVPR'22 Real 39k - ✔ - 2 ✔ ✔ - ✔ - - - Link
V2XSet [5] ECCV'22 Simu 11K ✔ ✔ - 2-5 ✔ ✔ - ✔ - - - Link
DOLPHINS [6] ACCV'22 Simu 42k ✔ ✔ - 3 ✔ ✔ - ✔ - - - Link
V2V4Real [7] CVPR'23 Real 20K ✔ - - 2 ✔ ✔ - ✔ - ✔ - Link
V2X-Seq [8] CVPR'23 Real 15k - ✔ - 2 ✔ ✔ - ✔ - ✔ ✔ Link
DeepAccident [9] AAAI'24 Simu 57K ✔ ✔ - 1-5 ✔ ✔ - ✔ ✔ ✔ ✔ Link
HoloVIC [10] CVPR'24 Real 100K - ✔ - 2 ✔ ✔ - ✔ - ✔ - Link
TUMTraf-V2X [11] CVPR'24 Real 2K - ✔ - 2 ✔ ✔ - ✔ - ✔ - Link
R-Cooper [12] CVPR'24 Real 30K - - ✔ 2 ✔ ✔ - ✔ - ✔ - Link
MARS [13] CVPR'24 Real 1.5k ✔ - - 2-3 ✔ ✔ - - - - - Link
V2X-Real [14] ECCV'24 Real 33k ✔ ✔ ✔ 2-4 ✔ ✔ - ✔ - - - Link

Notes:

  • Source: simulator (Simu) and real-world (Real).
  • Frame refers to annotated LiDAR-based cooperative perception frame number.
  • Supported common perception tasks: 3D object detection (OD), BEV semantic segmentation (SS), 3D object tracking (OT), motion prediction (MP).

Back to Contents 🔙

References:

  1. V2VNet: Vehicle-to-Vehicle Communication for Joint Perception and Prediction (ECCV'20)
  2. V2X-Sim: Multi-Agent Collaborative Perception Dataset and Benchmark for Autonomous Driving (RAL'21)
  3. OPV2V: An open benchmark dataset and fusion pipeline for perception with vehicle-to-vehicle communication (ICRA'22)
  4. DAIR-V2X: A Large-Scale Dataset for Vehicle-Infrastructure Cooperative 3D Object Detection (CVPR'22)
  5. V2X-ViT: Vehicle-to-everything cooperative perception with vision transformer (ECCV'22)
  6. DOLPHINS: Dataset for Collaborative Perception enabled Harmonious and Interconnected Self-driving (ACCV'22)
  7. V2V4Real: A Real-world Large-scale Dataset for Vehicle-to-Vehicle Cooperative Perception (CVPR'23)
  8. V2X-Seq: A large-scale sequential dataset for vehicle-infrastructure cooperative perception and forecasting (CVPR'23)
  9. DeepAccident: A Motion and Accident Prediction Benchmark for V2X Autonomous Driving (AAAI'24)
  10. HoloVIC: Large-scale Dataset and Benchmark for Multi-Sensor Holographic Intersection and Vehicle-Infrastructure Cooperative (CVPR'24)
  11. TUMTraf V2X Cooperative Perception Dataset (CVPR'24)
  12. RCooper: A Real-world Large-scale Dataset for Roadside Cooperative Perception (CVPR'24)
  13. Multiagent Multitraversal Multimodal Self-Driving: Open MARS Dataset (CVPR'24)
  14. V2X-Real: a Largs-Scale Dataset for Vehicle-to-Everything Cooperative Perception (ECCV'24)