Image Evolution GIF Generator using GA

A genetic algorithm implementation that evolves a low-resolution image by attempting to approximate a target image through evolutionary computation. The program generates an animated GIF that visualizes the entire evolution process from random noise to a recognizable approximation using pixel-level mutations and selection.

Demo

Original Image	Generated GIF

Details

This project uses a genetic algorithm to evolve RGB pixel values for each position in a 100x100 image grid. Each pixel is represented by a chromosome containing binary genes for red, green, and blue color channels. Through selection, crossover, and mutation operations, the algorithm iteratively improves the image quality to match a target image.

Genetic Algorithm Parameters

const IMG_SIZE: usize = 100;           // Image dimensions (100x100)
const POPULATION_SIZE: usize = 6;      // Population size per pixel
const ITERATION: usize = 50;           // Number of generations
const MUTATION_RATE: f64 = 0.05;       // Bit-flip mutation probability
const CROSSOVER_RATE: f64 = 0.8;       // Crossover probability
const GENE_LENGTH: usize = 8;          // Bits per color channel
const RGB_CHANNELS: usize = 3;         // Red, Green, Blue channels
const TOURNAMENT_SIZE: usize = 3;      // Tournament selection size
const ELITE_SIZE: usize = 2;           // Number of elite individuals preserved

For more details, see here.

Usage

Prerequisites

Add these dependencies to your Cargo.toml:

[dependencies]
image = "0.24"
rand = "0.8"
gif = "0.12"

Running the Program

# Place a target image named target.png in the project directory.
cargo run

Output Files

• result.png: Final evolved image
• result.gif: Animated evolution process
• target_sample.png: Copy of the target image used