# BDNet: Bengali Handwritten Numeral Digit Recognition based on Densely connected Convolutional Neural Networks

## TL;DR

BDNet is a densely connected CNN model that achieves high accuracy in recognizing Bengali handwritten digits, significantly reducing error rates compared to previous models, and is trained with innovative data augmentation techniques.

## Contribution

The paper introduces BDNet, a novel densely connected CNN architecture for Bengali digit recognition, with improved accuracy and robustness through unique data preprocessing and augmentation.

## Key findings

- Achieved 99.775% test accuracy on Bengali handwritten digits
- Reduced error rate by 62.5% compared to previous models
- Created a new dataset of 1000 Bengali handwritten numeral images

## Abstract

Images of handwritten digits are different from natural images as the orientation of a digit, as well as similarity of features of different digits, makes confusion. On the other hand, deep convolutional neural networks are achieving huge success in computer vision problems, especially in image classification. BDNet is a densely connected deep convolutional neural network model used to classify (recognize) Bengali handwritten numeral digits. It is end-to-end trained using ISI Bengali handwritten numeral dataset. During training, untraditional data preprocessing and augmentation techniques are used so that the trained model works on a different dataset. The model has achieved the test accuracy of 99.775%(baseline was 99.40%) on the test dataset of ISI Bengali handwritten numerals. So, the BDNet model gives 62.5% error reduction compared to previous state-of-the-art models. Here we have also created a dataset of 1000 images of Bengali handwritten numerals to test the trained model, and it giving promising results. Codes, trained model and our own dataset are available at: {https://github.com/Sufianlab/BDNet}.