BHDD: A Burmese Handwritten Digit Dataset

Swan Htet Aung; Hein Htet; Htoo Say Wah Khaing; Thuya Myo Nyunt

cs.CVMar 23, 2026

BHDD: A Burmese Handwritten Digit Dataset

Authors: Swan Htet Aung, Hein Htet, Htoo Say Wah Khaing, Thuya Myo Nyunt

Organizations: Expa.AI Research Team, Yangon, Myanmar

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

Figures & tables

Fig. 1: Sample collection sheets on plain paper (left) and grid paper (middle two), and the mobile preprocessing app (right) that let contributors verify digit extraction before submission.

Fig. 2: Samples from all ten Burmese digit classes (rows: 0–9).

Fig. 3: Training set (balanced, 6,000 per class) vs. test set (unbalanced, 6,856 down to 389).

Fig. 4: Pixel intensity distributions per class. Left: log-scale histogram. Right: per-class kernel density estimates.

Fig. 5: Mean image per class (6,000 training samples each). Sharp shapes show consistent digit forms across writers.

Fig. 6: Per-class pixel variance. Bright areas mark where writers differ most.

Fig. 7: Writing style variation within classes 0, 3, 5, and 8.

Fig. 8: Commonly confused digit pairs with error counts from the improved CNN.

Model	Accuracy	F1	Precision	Recall
MLP	0.9940	0.9905	0.9876	0.9934
CNN	0.9975	0.9964	0.9959	0.9970
Improved CNN	0.9983	0.9980	0.9972	0.9988

TABLE I: Baseline Results on the BHDD Test Set

Fig. 9: Improved CNN confusion matrix. Nearly all off-diagonal entries are zero. The 0–1 pair dominates the remaining errors.

Explore similar work

myMNIST: Benchmark of PETNN, KAN, and Classical Deep Learning Models for Burmese Handwritten Digit Recognition
Mar 19, 2026Ye Kyaw Thu, Thazin Myint Oo, Thepchai SupnithiOptical Character Recognition Kolmogorov-Arnold Networks
National Electronics and Computer Technology Center (NECTEC), Pathumthani, Thailand · Language Understanding Lab., Yangon, Myanmar
Multi-Head Attention based interaction-aware architecture for Bangla Handwritten Character Recognition: Introducing a Primary Dataset
Apr 8, 2026Mirza Raquib, Asif Pervez Polok, Kedar Nath Biswas +3Optical Character Recognition Hybrid CNN-Transformer Architectures
Department of Computer and Communication Engineering, International Islamic University Chittagong, Bangladesh. · mPower Social Enterprise. · Department of Information and Communication Engineering, Noakhali Science and Technology University, Bangladesh. +2
DohaScript: A Large-Scale Multi-Writer Dataset for Continuous Handwritten Hindi Text
Feb 20, 2026Kunwar Arpit Singh, Ankush Prakash, Haroon R LoneOptical Character Recognition
IISER Bhopal, Bhopal, India

Mar 19, 2026J/K move · Enter open · S save

myMNIST: Benchmark of PETNN, KAN, and Classical Deep Learning Models for Burmese Handwritten Digit Recognition

Ye Kyaw Thu, Thazin Myint Oo, Thepchai Supnithi

National Electronics and Computer Technology Center (NECTEC), Pathumthani, Thailand · Language Understanding Lab., Yangon, Myanmar

Optical Character Recognition Kolmogorov-Arnold Networks

Open paper AskPDF arXiv

We present the first systematic benchmark on a standardized iteration of the publicly available Burmese Handwritten Digit Dataset (BHDD), which we have designated as myMNIST Benchmarking. While BHDD serves as a foundational resource for Myanmar NLP/AI, it lacks a comprehensive, reproducible performance baseline across modern architectures. We evaluate eleven architectures spanning classical deep learning models (Multi-Layer Perceptron, Convolutional Neural Network, Long Short-Term Memory, Gated Recurrent Unit, Transformer), recent alternatives (FastKAN, EfficientKAN), an energy-based model (JEM), and physics-inspired PETNN variants (Sigmoid, GELU, SiLU). Using Precision, Recall, F1-Score, and Accuracy as evaluation metrics, our results show that the CNN remains a strong baseline, achieving the best overall scores (F1 = 0.9959, Accuracy = 0.9970). The PETNN (GELU) model closely follows (F1 = 0.9955, Accuracy = 0.9966), outperforming LSTM, GRU, Transformer, and KAN variants. JEM, representing energy-based modeling, performs competitively (F1 = 0.9944, Accuracy = 0.9958). KAN-based models (FastKAN, EfficientKAN) trail the top performers but provide a meaningful alternative baseline (Accuracy ~0.992). These findings (i) establish reproducible baselines for BHDD across diverse modeling paradigms, (ii) highlight PETNN's strong performance relative to classical and Transformer-based models, and (iii) quantify the gap between energy-inspired PETNNs and a true energy-based model (JEM). We release this benchmark to facilitate future research on Myanmar digit recognition and to encourage broader evaluation of emerging architectures on regional scripts.

Abstract

Explore similar work

myMNIST: Benchmark of PETNN, KAN, and Classical Deep Learning Models for Burmese Handwritten Digit Recognition

Multi-Head Attention based interaction-aware architecture for Bangla Handwritten Character Recognition: Introducing a Primary Dataset

DohaScript: A Large-Scale Multi-Writer Dataset for Continuous Handwritten Hindi Text