Persian-Digits-OCR

(Handwrite)

neural network "DeepLearning"

This project focuses on recognizing Persian digits using a deep learning-based Optical Character Recognition (OCR) system. The model has been trained on a custom dataset of Persian digits, and the pre-trained model is available for immediate use.

Table of Contents

• Installation
• Usage
• Project Structure
• Dataset
• Pre-Trained Model
• License

Installation

1. Clone the repository:

   git clone https://github.com/tahabakhtari/Persian-digits-ocr.git

2. Install required dependencies:

   pip install -r requirements.txt

Usage

Run the Jupyter notebook to perform Persian digit recognition:

1. Load the Persian digit dataset.
2. Preprocess the dataset to fit the model's input requirements.
3. Use the pre-trained model for predictions or train a new model.
4. Evaluate the model's performance on test data.

jupyter notebook code.ipynb

Using the Pre-Trained Model

To use the pre-trained model, load the model file and run predictions directly on your input images:

from keras.models import load_model

model = load_model('models/persian_digit_model.h5')
# Use the model for predictions
predictions = model.predict(input_data)

Project Structure

• code.ipynb: The core notebook that includes data loading, model training, and prediction steps.
• data/: Directory containing the dataset of Persian digit images.
• models/: Contains the pre-trained model (persian_digit_model.h5).
• utils.py: Helper functions for data preprocessing and evaluation.

Dataset

The dataset consists of Persian digit images (0-9) that have been preprocessed and formatted for model training. These images are available in the data/ folder, and can be directly used to retrain or fine-tune the model.

Pre-Trained Model

A pre-trained model is included in the models/ directory. You can load and use this model without needing to retrain. The model has been trained on thousands of Persian digit samples to ensure accurate recognition.

License

This project is licensed under the MIT License - see the LICENSE file for details.