Iranis

Category: Datasets
Tags: datasets, computer-vision, transportation

A large-scale dataset of Farsi license plate characters (numbers, letters, and symbols) extracted from real-world license plate images captured by various cameras. The dataset is designed for character recognition and license plate recognition research.

Features

• Scale and content

  • More than 83,000 labeled character images (total 83,844 instances)
  • Characters cropped from real-world Farsi license plates
  • Includes numbers, letters, and a special symbol category

• Character categories and class distribution

  • Numbers

    • 0: 2,501 instances
    • 1: 3,495 instances
    • 2: 3,930 instances
    • 3: 2,745 instances
    • 4: 5,774 instances
    • 5: 3,610 instances
    • 6: 5,753 instances
    • 7: 3,736 instances
    • 8: 3,583 instances
    • 9: 3,528 instances

  • Letters (Farsi characters with Latin label names)

    • A (الف): 2,517 instances
    • B (ب): 2,511 instances
    • P (پ): 2,519 instances
    • J (ج): 2,505 instances
    • H (ه): 2,558 instances
    • D (د): 2,504 instances
    • Sin (س): 2,445 instances
    • Sad (ص): 2,515 instances
    • T (ط): 2,512 instances
    • Gh (ق): 2,482 instances
    • L (ل): 2,502 instances
    • M (م): 2,500 instances
    • N (ن): 2,558 instances
    • V (و): 2,509 instances
    • Y (ی): 2,491 instances
    • PuV (Public Vehicle, ع): 2,508 instances
    • Taxi (ت): 2,551 instances

  • Symbol

    • PwD (People with Disabilities, 🦽): 2,502 instances

• Total size

  • Overall total: 83,844 labeled character images

• Intended use

  • Farsi license plate character recognition
  • License plate recognition (LPR) systems
  • Pattern recognition and computer vision research

• Availability

  • Primary project page: https://alitourani.github.io/Iranis-dataset/
  • Also mirrored on:
    • Kaggle: https://www.kaggle.com/sajjadsoroori/iranis
    • Data.world: https://data.world/guilanits/iranis

Citation

If you use the Iranis dataset in academic or research work, the authors request that you cite:

@inproceedings{Tourani2021,
    author = {Tourani, Ali and Soroori, Sajjad and Shahbahrami, Asadollah and Akoushideh, Alireza},
    title = {},
    booktitle = {2021 5th International Conference on Pattern Recognition and Image Analysis (IPRIA)},
    doi = {10.1109/IPRIA53572.2021.9483461},
    isbn = {978-1-6654-2659-6},
    pages = {1--5},
    publisher = {IEEE},
    url = {https://ieeexplore.ieee.org/document/9483461/},
    year = {2021}
}

Pricing

• Not specified; dataset is distributed via the project page and public data platforms (Kaggle, Data.world). Usage terms and any licensing or access restrictions should be checked on the respective hosting pages.