Isolated Bangla Handwritten Character Classification using Transfer Learning

2025-09-03

AI Summary

This study addresses the fine-grained classification of isolated handwritten Bangla characters—comprising 50 basic and 34 compound classes—under limited-sample conditions. We propose a transfer learning–based multi-model ensemble deep framework that innovatively integrates 3D convolutional neural networks (3DCNNs) with lightweight backbone architectures (ResNet, MobileNet). To mitigate gradient vanishing and enhance discriminability for low-resource compound characters, we introduce feature recalibration and cross-layer residual connections. Evaluated on the standard Bangla Lekha Isolated dataset, our model achieves 99.82% accuracy on the training set and 99.46% on the test set—substantially surpassing existing state-of-the-art methods. The results demonstrate the framework’s effectiveness and generalizability in recognizing regionally specific scripts characterized by structural complexity and high inter-class visual similarity.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Natural Language Processing: Learning & Optimization for NLP
• Computer Vision: Generative Adversarial Networks (GANs) for Vision

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterization
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Bangla language consists of fifty distinct characters and many compound characters. Several notable studies have been performed to recognize Bangla characters, both handwritten and optical. Our approach uses transfer learning to classify the basic, distinct, as well as compound Bangla handwritten characters while avoiding the vanishing gradient problem. Deep Neural Network techniques such as 3D Convolutional Neural Network (3DCNN), Residual Neural Network (ResNet), and MobileNet are applied to generate an end-to-end classification of all possible standard formations of handwritten characters in the Bangla language. The Bangla Lekha Isolated dataset, which contains 166,105 Bangla character image samples categorized into 84 distinct classes, is used for this classification model. The model achieved 99.82% accuracy on training data and 99.46% accuracy on test data. Comparisons with various state-of-the-art benchmarks of Bangla handwritten character classification show that the proposed model achieves better accuracy in classifying the data.

Problem

• Classifying isolated Bangla handwritten characters using transfer learning
• Addressing vanishing gradient problem in deep neural networks
• Achieving high accuracy across basic and compound character types

Innovation

• Transfer learning for Bangla character classification
• 3DCNN, ResNet, MobileNet deep neural networks
• End-to-end classification avoiding vanishing gradient