- BaiduYun Drive password:fp52
- DropBox
LSV-LP
├── test
│ ├── jsons
│ │ ├── move2move
│ │ │ ├── 1
│ │ │ │ ├── frame1.json
│ │ │ │ ├── ...
│ │ ├── move2static
│ │ └── static2move
│ └── videos
│ ├── move2move
│ │ ├── 1
│ │ │ ├── frame1.jpg
│ │ │ ├── ...
│ ├── move2static
│ └── static2move
├── train
│ ├── jsons
│ │ ├── move2move
│ │ ├── move2static
│ │ └── static2move
│ └── videos
│ ├── move2move
│ ├── move2static
│ └── static2move
└── validation
├── jsons
│ ├── move2move
│ ├── move2static
│ └── static2move
└── videos
├── move2move
├── move2static
└── static2move
- 'id': the id of the car and its plate in this video;
- 'carBox': the bounding box of the car in this frame;
- 'licPoly': the coordinates of the four vertices of the current license plate;
- 'licNumber': the number of the current license plate;
Please cite the paper 《Q. Wang, X. Lu, C. Zhang, Y. Yuan and X. Li, "LSV-LP:Large-Scale Video-Based License Plate Detection and Recognition," IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI), 2022》, if you benefit from this dataset.
It is worth noting that LSV-LP can only be used for academic research, not for commercial use.