endy-alunge/ZAR-Bank-Notes-Recognition
Record type:
software
Creator:
end
Host:
A deep learning and computer vision framework for South African Banknote (ZAR) detection. Built with Flask, application processes uploaded currency scans through custom preprocessing, region segmentation, and feature extraction stages to accurately classify denominations and verify note authenticity in real time.
# ZAR Bank Notes Recognition
A deep learning and computer vision framework for South African Banknote (ZAR) detection and authentication.
The system processes uploaded banknote images through a custom image processing pipeline consisting of preprocessing, region segmentation, feature extraction, and machine learning classification. The application is designed to identify South African currency denominations and assist in verifying note authenticity in real time.
---
## GitHub Repository
Repository URL:
github.com
---
## Features
* South African banknote denomination recognition
* Image preprocessing and enhancement
* Region-based segmentation
* Feature extraction using computer vision techniques
* Machine learning classification
* Real-time prediction through a Flask web application
* Model persistence and reuse
* User-friendly web interface
---
## Technologies Used
* Python
* OpenCV
* NumPy
* Scikit-Image
* Scikit-Learn
* Flask
* Matplotlib
* Seaborn
* Pillow
---
## Requirements
Install the following dependencies:
opencv-python>=4.8.1.78
numpy>=1.26.0
scikit-image>=0.21.0
scikit-learn>=1.3.0
matplotlib>=3.7.2
seaborn>=0.12.2
flask>=2.3.3
pillow>=10.3.0
joblib>=1.3.2
Alternatively, install all dependencies from a requirements file:
bash
pip install -r requirements.txt
---
## Architecture Overview
The application follows a layered architecture:
1. Data Layer
* Stores raw, processed, and testing banknote images.
2. Backend Layer
* Performs preprocessing, segmentation, feature extraction, and classification.
* The pipeline orchestrates the complete recognition workflow.
3. Model Layer
* Stores trained machine learning models used for denomination prediction and authenticity analysis.
4. Frontend Layer
* Provides a web-based interface for image uploads and displaying prediction results.
5. Application Entry Points
* `main.py` trains and saves machine learning models.
* `app.py` launches the …