BHDD: A Burmese Handwritten Digit Dataset

TL;DR

The paper introduces BHDD, a large dataset of handwritten Burmese digits with analysis and baseline models, facilitating research in Burmese OCR and handwritten digit recognition.

Contribution

It presents the first extensive Burmese handwritten digit dataset with detailed analysis and baseline models, supporting future research in this language's OCR.

Findings

Baseline models achieve over 99.8% accuracy.

Dataset includes diverse samples from over 150 contributors.

Analysis reveals common confusions among similar-shaped digits.

Abstract

We introduce the Burmese Handwritten Digit Dataset (BHDD), a collection of 87,561 grayscale images of handwritten Burmese digits in ten classes. Each image is 28x28 pixels, following the MNIST format. The training set has 60,000 samples split evenly across classes; the test set has 27,561 samples with class frequencies as they arose during collection. Over 150 people of different ages and backgrounds contributed samples. We analyze the dataset's class distribution, pixel statistics, and morphological variation, and identify digit pairs that are easily confused due to the round shapes of the Myanmar script. Simple baselines (an MLP, a two-layer CNN, and an improved CNN with batch normalization and augmentation) reach 99.40%, 99.75%, and 99.83% test accuracy respectively. BHDD is available under CC BY-SA 4.0 at https://github.com/baseresearch/BHDD