SEACrowd/kvis_th_ocr

The KVIS Thai OCR Dataset contains scanned handwritten version of all 44 Thai characters obtained from 27 individuals. It consisted of 1079 images from 44 classes (letters). This dataset consists of all Thai consonants with different writing styles of various people from ages between 16 and 75. Vowels and intonation are not taken into consideration for the dataset collected.

Languages

tha

Supported Tasks

Optical Character Recognition

Dataset Usage

Using datasets library

from datasets import load_dataset
dset = datasets.load_dataset("SEACrowd/kvis_th_ocr", trust_remote_code=True)

Using seacrowd library

# Load the dataset using the default config
dset = sc.load_dataset("kvis_th_ocr", schema="seacrowd")
# Check all available subsets (config names) of the dataset
print(sc.available_config_names("kvis_th_ocr"))
# Load the dataset using a specific config
dset = sc.load_dataset_by_config_name(config_name="<config_name>")

More details on how to load the seacrowd library can be found here.

Dataset Homepage

https://data.mendeley.com/datasets/8nr3pbdk5c/1

Dataset Version

Source: 1.0.0. SEACrowd: 2024.06.20.

Dataset License

Creative Commons Attribution 4.0 (cc-by-4.0)

Citation

If you are using the Kvis Th Ocr dataloader in your work, please cite the following:

@INPROCEEDINGS{8584876,
  author={Joseph, Ferdin Joe John and Anantaprayoon, Panatchakorn},
  booktitle={2018 International Conference on Information Technology (InCIT)},
  title={Offline Handwritten Thai Character Recognition Using Single Tier Classifier and Local Features},
  year={2018},
  volume={},
  number={},
  pages={1-4},
  abstract={Handwritten character recognition is a conversion process of handwriting into machine-encoded text. Currently,
  several techniques and methods are proposed to enhance accuracy of handwritten character recognition for many languages
  spoken across the globe. In this project, a local feature-based approach is proposed to enhance the accuracy of handwritten
  offline character recognition for Thai alphabets. In the experiment, through MATLAB, 100 images for each class of Thai
  alphabets are collected and k-fold cross validation is applied to manage datasets to train and test. A gradient invariant
  feature set consisting of LBP and shape features is extracted. The classification is operated by using query matching based
  on Euclidean distance. The accuracy would be the percentage of correct classification for each class. For the result, the
  highest accuracy is 68.96% which has 144-bit shape features and uniform pattern LBP for the features.},
  keywords={Character recognition;Feature extraction;Shape;Genetic algorithms;Matlab;Gray-scale;Optical character recognition
  software;Offline Character Recognition;Local Binary Pattern;Thai Handwriting},
  doi={10.23919/INCIT.2018.8584876},
  ISSN={},
  month={Oct},
  url={https://ieeexplore.ieee.org/document/8584876}}



@article{lovenia2024seacrowd,
    title={SEACrowd: A Multilingual Multimodal Data Hub and Benchmark Suite for Southeast Asian Languages}, 
    author={Holy Lovenia and Rahmad Mahendra and Salsabil Maulana Akbar and Lester James V. Miranda and Jennifer Santoso and Elyanah Aco and Akhdan Fadhilah and Jonibek Mansurov and Joseph Marvin Imperial and Onno P. Kampman and Joel Ruben Antony Moniz and Muhammad Ravi Shulthan Habibi and Frederikus Hudi and Railey Montalan and Ryan Ignatius and Joanito Agili Lopo and William Nixon and Börje F. Karlsson and James Jaya and Ryandito Diandaru and Yuze Gao and Patrick Amadeus and Bin Wang and Jan Christian Blaise Cruz and Chenxi Whitehouse and Ivan Halim Parmonangan and Maria Khelli and Wenyu Zhang and Lucky Susanto and Reynard Adha Ryanda and Sonny Lazuardi Hermawan and Dan John Velasco and Muhammad Dehan Al Kautsar and Willy Fitra Hendria and Yasmin Moslem and Noah Flynn and Muhammad Farid Adilazuarda and Haochen Li and Johanes Lee and R. Damanhuri and Shuo Sun and Muhammad Reza Qorib and Amirbek Djanibekov and Wei Qi Leong and Quyet V. Do and Niklas Muennighoff and Tanrada Pansuwan and Ilham Firdausi Putra and Yan Xu and Ngee Chia Tai and Ayu Purwarianti and Sebastian Ruder and William Tjhi and Peerat Limkonchotiwat and Alham Fikri Aji and Sedrick Keh and Genta Indra Winata and Ruochen Zhang and Fajri Koto and Zheng-Xin Yong and Samuel Cahyawijaya},
    year={2024},
    eprint={2406.10118},
    journal={arXiv preprint arXiv: 2406.10118}
}