💵 PKR Note Recognizer

🌐 Live Demo

👉 Click here to try the PKR Note Recognizer

An AI-powered web application to recognize Pakistani currency notes using a deep learning model (MobileNetV2), a Flask backend, and a responsive web interface.

---

🔗 Related Repositories

• 🔧 Backend (Flask + MobileNetV2 Model)
  pkr_note_recognizer_backend

• 🎨 Frontend (Web Interface)
  pkr_note_recognizer_frontend

---

🎯 Real-World Problem It Solves

Accurately identifying currency notes is critical in many everyday situations:

• Visually impaired individuals struggle to recognize note denominations.
• Businesses face human error in handling and verifying cash.
• Financial institutions require fast, automated note recognition for machines like ATMs, deposit kiosks, or cash counters.

This system provides an AI-based solution to recognize Pakistani Rupee (PKR) notes in real-time from images. It can be integrated into:

• Mobile apps for accessibility
• Retail checkout systems
• Currency validation hardware
• Educational or financial tech tools

---

🧠 Overview

The PKR Note Recognizer is a full-stack machine learning project that allows users to upload images of PKR notes and get the denomination prediction instantly. The backend uses a fine-tuned MobileNetV2 model, and the frontend provides a simple, intuitive interface.

---

📸 Demo

---

⚙️ Tech Stack

Layer	Technology
Frontend	React
Backend	Flask (Python)
Model	MobileNetV2 (TensorFlow/Keras)
Deployment	Render,Vercel

---

🧩 Architecture

[ User ]
   ↓
[ Frontend ]
   ↓ HTTP (Image Upload)
[ Flask API ]
   ↓
[ Trained MobileNetV2 Model ]
   ↓
[ Predicted Note Denomination ]

Setup Instructions

✅ Prerequisites

• Python 3.8+
• Node.js and npm (if using React frontend)
• Git

Clone the Main Project (with Submodules)

git clone --recurse-submodules https://github.com/yourusername/pkr_note_recognizer.git
cd pkr_note_recognizer

If you forgot --recurse-submodules, run:

git submodule update --init --recursive

Backend Setup (Flask + Model)

cd backend
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
pip install -r requirements.txt

# Run the server
python app.py

• Backend runs at: http://localhost:5000
• API endpoint: POST /predict (accepts image file)

Frontend Setup (React)

cd ../frontend
npm install
npm start

• Frontend runs at: http://localhost:3000
• It connects to the Flask backend to fetch predictions.

Dataset & Model

• Dataset: Custom-collected images of PKR notes (10, 20, 50, 100, 500, 1000, 5000)
• Model: MobileNetV2 pretrained on ImageNet, fine-tuned on this dataset
• Accuracy: ~[90+]% on test data

Features

• Upload image of PKR note and get prediction instantly
• Lightweight model for fast inference
• Simple and responsive web UI
• Clean API interface for future integrations

Future Improvements

• Deploy as mobile app using React Native / Flutter
• Add audio output for accessibility
• Support multiple currencies -Improve model with larger dataset and augmentations
• Add history/log of predictions

🧑‍💻 Author Aayan Ali Hashim GitHub LinkedIn