aungthuhein-dev/burmese-character-dataset

Burmese Character Dataset (Combined)

A combined image-classification dataset of handwritten and air-written Burmese (Myanmar) characters and digits, merged from two publicly available Kaggle datasets into a single imagefolder-compatible layout (one subfolder per character class).

• Classes: 54 (digits 0–9 plus Burmese consonant/vowel characters, e.g. Ah, Aha, O, ba_kone, ga_nge, nga, za_kwear, ...)
• Images: 13,269 JPG files
• Size: ~106 MB

Source datasets & attribution

This dataset is a derivative work combining the following two Kaggle datasets. All credit for the original data collection belongs to their respective authors.

1. Hand-Written Burmese Characters And Digits Author: DataScience&MLClub Source: https://www.kaggle.com/datasets/datasciencemlclub/burmesecharactersanddigit License: CC BY-SA 4.0 Description (from source): Handwritten data collected from 223 people, each writing 54 Myanmar characters, totaling at least 12,042 handwritten images.

2. Air-written Burmese Alphabets and Numerals Dataset Author: Banyar Oo Source: https://www.kaggle.com/datasets/banyaroo26/air-written-burmese-alphabets License: CC BY 4.0 Description (from source): Air-written Burmese alphabets and numerals, primarily 83×84 pixels, 30 images per character (some at 63×64).

License

This combined dataset is released under CC BY-SA 4.0 (https://creativecommons.org/licenses/by-sa/4.0/).

CC BY-SA 4.0 is used because one of the source datasets (#1) is itself licensed CC BY-SA 4.0, which requires any adaptation or combination built from it to be shared under the same license. CC BY 4.0 material (#2) is compatible with, and may be incorporated into, a CC BY-SA 4.0 work.

If you use or redistribute this dataset:

• Give attribution to both original authors/sources listed above.
• Any derivative you create from this dataset must also be licensed under CC BY-SA 4.0 (or a CC-listed compatible license) — this is the ShareAlike requirement inherited from source #1.

Dataset structure

<class_name>/
  <image>.jpg
  ...

Each top-level folder name is the class label. Loadable directly with 🤗 datasets as an imagefolder:

from datasets import load_dataset

dataset = load_dataset("aungthuhein-dev/burmese-character-dataset")

Citation

If you use this dataset, please cite both original sources:

DataScience&MLClub. "Hand-Written Burmese Characters And Digits." Kaggle.
https://www.kaggle.com/datasets/datasciencemlclub/burmesecharactersanddigit

Banyar Oo. "Air-written Burmese Alphabets and Numerals Dataset." Kaggle.
https://www.kaggle.com/datasets/banyaroo26/air-written-burmese-alphabets