README.md

metadata

dataset_info:
  features:
    - name: image
      dtype: image
    - name: response
      dtype: string
  splits:
    - name: train
      num_bytes: 41407347
      num_examples: 228
  download_size: 41405508
  dataset_size: 41407347
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: apache-2.0
task_categories:
  - image-to-text
language:
  - en
size_categories:
  - n<1K

Number-Plate-Recognition

A lightweight dataset for automatic number plate recognition (ANPR). Each sample contains a vehicle image paired with the vehicle color and the corresponding license plate number in a standardized JSON format.

Features

• Vehicle images
• Vehicle color annotation
• License plate text (OCR)
• JSON formatted labels
• Suitable for OCR and vision-language model training

Label Format

{
  "vehicle_color": "black",
  "license_plate_number": "77-ZFK-8"
}

Use Cases

• Automatic Number Plate Recognition (ANPR)
• OCR model training and evaluation
• Vision-language model fine-tuning
• Vehicle attribute extraction
• Multimodal benchmarking

License

Apache-2.0