This repository contains code for building an Optical Character Recognition (OCR) system specifically designed for Persian alphabets. The system utilizes the LeNet5 architecture, a classic convolutional neural network (CNN) known for its effectiveness in image recognition tasks.
The OCR system is trained on a dataset comprising approximately 570,000 labeled images of Persian alphabets. This dataset is publicly available on Kaggle and can be accessed here.
After training on the provided dataset, the OCR system achieves an accuracy of around 91% on classification tasks.
The LeNet5 architecture is chosen for its simplicity and effectiveness in handling image classification tasks. It consists of multiple layers including convolutional, pooling, and fully connected layers, providing a robust framework for recognizing patterns within images.
The code provided in this repository can be used to train and evaluate the OCR system on new data. Additionally, pretrained models and scripts for inference are included for easy deployment.