Bangla Handwritten Character Recognition

---

Overview

This project implements a Convolutional Neural Network (CNN) to recognize handwritten Bangla characters using the BanglaLekha-Isolated dataset.
It classifies 84 characters including vowels, consonants, and digits. The notebook covers data preprocessing, model training, evaluation, and visualization of predictions with Bangla labels.

---

Dataset

• Name: BanglaLekha-Isolated
• Source: Dataset
• Contains images of isolated handwritten Bangla characters.
• Labels are mapped from 1–84.

---

Installation

Clone the repository:

git clone https://github.com/utpalbarua/Bangla-Handwritten-Character-Recognition-Using-Deep-Learning.git

cd banglalekha-handwritten-recognition

Install required dependencies:

pip install -r requirements.txt

Dependencies

• Python 3.10+
• PyTorch
• Torchvision
• Matplotlib
• scikit-learn
• seaborn
• PIL

---

Acknowledgements

• BanglaLekha-Isolated dataset