Genetic Algorithm for Biometric Image Encryption Optimization

This project focuses on optimizing the parameters of a genetic algorithm (GA) to generate encrypted biometric images. The aim is to create cancelable biometric templates with high accuracy by fine-tuning the GA's crossover and mutation rates.

Table of Contents

• Project Overview
• Prerequisites
• Installation
• Usage
• Project Structure
• Functions Description
• Acknowledgements

Project Overview

The project involves the following steps:

1. Generate Initial Population: Create an initial population of encrypted biometric images using multiple rounds of XOR encryption with random keys.
2. Optimize GA Parameters: Use the genetic algorithm to optimize the crossover and mutation rates for generating cancelable biometric templates.
3. Evaluate Accuracy: Evaluate the accuracy of the cancelable templates generated by the GA.

Prerequisites

• Python 3.x
• OpenCV
• NumPy

Installation

1. Clone the repository:

   git clone https://github.com/JayArcher9/TRWI.git
   cd TRWI

2. Install the required Python packages:

   pip install opencv-python numpy

Usage

1. Place the original biometric image in the project directory. The image should be named download.jpg.
2. Run the main script:

   python main.py

3. The script will generate initial encrypted images and optimize the GA parameters, printing the accuracy scores for each combination of crossover and mutation rates.

Project Structure

• main.py: Main script to execute the project.
• generate_initial_population: Function to generate initial population images using multiple rounds of encryption.
• optimize_ga_parameters: Function to optimize GA parameters (crossover and mutation rates).
• evaluate_population, select_parents, crossover_and_mutate, crossover, mutate, next_generation: Helper functions for GA operations.

Functions Description

generate_initial_population(original_biometric_image, num_images, num_rounds)

Generates an initial population of encrypted biometric images using XOR encryption.

Parameters:

• original_biometric_image: The original biometric image.
• num_images: Number of initial population images to generate.
• num_rounds: Number of encryption rounds.

Returns:

• initial_population: List of encrypted images.

optimize_ga_parameters(initial_population)

Optimizes the genetic algorithm parameters (crossover and mutation rates) to find the best cancelable template.

Parameters:

• initial_population: Initial population of encrypted images.

Returns:

• accuracy_scores: Dictionary containing accuracy scores for each combination of crossover and mutation rates.

Helper Functions

• evaluate_population(population): Evaluates the fitness of each cancelable template in the population.
• select_parents(population, fitness_scores): Selects parent cancelable templates for reproduction based on fitness scores.
• crossover_and_mutate(parents, crossover_rate, mutation_rate): Performs crossover and mutation operations on the parent cancelable templates.
• crossover(parent1, parent2): Performs crossover operation on two parent cancelable templates.
• mutate(individual, mutation_rate): Performs mutation operation on a cancelable template.
• next_generation(population, offspring, fitness_scores, offspring_fitness): Combines the offspring and parents, and selects the fittest individuals for the next generation.

Acknowledgements

This project was developed using Python, OpenCV, and NumPy. Special thanks to the open-source community for providing valuable resources and libraries.