tejasrisaladi/surge-antispoof-detection-using-ml-techniques

Raspberry Pi Fingerprint Sensor and Anti spoof Detection using ML techniques

This Python script is part of a research project conducted by Team X under the supervision of Professor Tushar Sandhan during the summer internship. The project focuses on integrating a fingerprint sensor with an RGB display using a Raspberry Pi, combining hardware interfacing, fingerprint operations, and machine learning for enhanced security.

Abstract

The project involves implementing a machine learning model designed to de- tect spoofing attempts, deployed on a Raspberry Pi hardware setup. The model analyze fingerprint images using innovative techniques like Ten Crop patch and crop processing. This approach not only expands the dataset but also enables the model to develop a keen eye for detail, effortlessly distinguishing genuine fingerprints from clever forgeries

Project Details

• Project Title: Integrating Fingerprint Sensor with RGB Display for Security Applications and Anti-Spoof Detection

• Proffessor: Professor Tushar Sandhan

• Project Team: Fingerprint Team

  • Tejasri
  • Sritan
  • Sanya
  • Prateek
  • Naina
  • Riddhima

• Mentors:

  • Aman
  • AjaySingh
  • Adarsh

Project directory Structure

SURGE/
│
├── README.md
|
├── src/
│   |__fingerprint_sensor_display.py
│   
├── images/
│   ├── image1.jpg
│   └── image2.png
│
└── requirements.txt

Requirements

Hardware Requirements

• Raspberry Pi: Tested on Raspberry Pi 3 Model B+
• Fingerprint Sensor: Compatible with PyFingerprint library, connected via UART (/dev/ttyUSB0)
• Adafruit RGB Display: ST7735R, connected via SPI

Software Requirements

• Python 3.x
• Python Libraries (Install via pip install -r requirements.txt):
  • RPi.GPIO: GPIO control for Raspberry Pi
  • Pillow: Image processing operations
  • pyfingerprint: Interface with the fingerprint sensor
  • torch, torchvision: Machine learning framework and models
  • adafruit-blinka: CircuitPython GPIO library compatibility layer

Development Environment

• Operating System: Recommended on Raspberry Pi OS (formerly Raspbian)
• Virtual Environment: Create and activate a virtual environment for Python dependencies.

Setup

Virtual Environment Setup

1. Create a Virtual Environment:

   • Open your terminal or command prompt.

   • Navigate to your project directory:

     cd path/to/your/project

   • Create a new virtual environment (replace venv with your preferred name):

     • For Windows:

       python -m venv venv

     • For macOS and Linux:

       python3 -m venv venv

2. Activate the Virtual Environment:

   • Activate the virtual environment:

     • For Windows:

       venv\Scripts\activate

     • For macOS and Linux:

       source venv/bin/activate

Installing Dependencies

3. Install Dependencies:

   • Install required Python libraries using pip with requirements.txt:

     pip install -r requirements.txt

4. Configure Hardware:

   • Connect the fingerprint sensor to the Raspberry Pi via UART (/dev/ttyUSB0).
   • Connect the Adafruit RGB display to the Raspberry Pi using SPI.

5. Configure Script:

   • Adjust GPIO pin numbers (BUTTON_PINS) and display pins (cs_pin, dc_pin, reset_pin, led_pin) in the Python script (fingerprint_sensor_display.py) based on your hardware setup.

6. Run Script:

   python fingerprint_sensor_display.py

Follow the instructions as displayed on the LCD.

Visuals

Complete Setup

Caption: Example caption describing the complete setup of the Raspberry Pi, fingerprint sensor, and RGB display integrated into a compact device.

Functionality

• Enroll Finger: Allows users to enroll new fingerprints into the system.
• Delete Finger: Deletes existing fingerprints from the sensor's database.
• Anti-Spoof Guard: Toggle feature to enable or disable spoof detection using machine learning.
• Search Finger: Searches for a fingerprint match in the system's database.(searches continuosly)
• Back:Return to main menu.

Functionality of Push Buttons

The Python script fingerprint_sensor_display.py utilizes push buttons connected to the Raspberry Pi GPIO pins for user interaction. Here's a brief overview of their functionality:

• Button 1: Enrolls a new fingerprint into the system.
• Button 2: Deletes an existing fingerprint from the sensor's database.
• Button 3: Toggles the anti-spoof guard feature, enabling or disabling spoof detection using machine learning.
• Button 4: Returns to the main menu after performing an operation or during idle states.

These buttons are configured using the RPi.GPIO library in the script. Each button press triggers specific functions related to fingerprint operations, display updates, and system configurations.

Dataset

• CrossMatch,Greenbit are used for cooperative framework.

Acknowledgement

We would like to express our sincere gratitude to Professor Tushar Sandhan for his invaluable guidance and support throughout this project. Special thanks to our mentors,Aman, Ajay Singh, and Adarsh, for their feedback which greatly contributed to the success of this project.

Notes

• Due to patent rights, the spoof detection algorithm's details and accuracy metrics cannot be disclosed.
• The compact device housing the entire setup was fabricated using 3D printing for robustness and integration.