GenAlgo

GenAlgo is a genetic algorithm-based image approximation project. This project utilizes semi-transparent triangles to approximate a target image and leverages GPU acceleration with CUDA for efficient computation of the fitness function.

Dependencies

• SFML 2.6
• CUDA 12.5

Installing Dependencies

Ubuntu

sudo apt update
sudo apt install libsfml-dev build-essential g++ nvidia-cuda-toolkit

Arch Linux

sudo pacman -Syu
sudo pacman -S cmake g++ gcc cuda sfml

Tested Hardware

• NVIDIA RTX 3060 Mobile
• NVIDIA GTX 1650 Mobile
• NVIDIA RTX 3060

Compiler

• GCC 14.1.1

Usage

To run GenAlgo, use the following command:

genalgo -i <image> [options]

Options

• -i, --input <image>: Input image file.
• -o, --output <svg>: Output SVG file.
• -gi, --gen-input <file>: Input file to continue from.
• -go, --gen-output <file>: Output file to save the generation.
• -s, --seed <seed>: Seed for the random number generator (default = platform-specific random).
• --period <n>: Number of generations between renders/logging (default = 50).
• --no-render: Disable rendering.
• --no-breed: Disable breeding.

Renderer Keybindings

• S: Toggle showing the original image.
• R: Reset to the best individual and update the display.
• N: Show the next individual in the population.
• P: Show the previous individual in the population.
• Up Arrow: Increase render scale.
• Down Arrow: Decrease render scale.

Building

To build GenAlgo, make sure you have all the dependencies installed and run the following commands:

mkdir build
cd build
cmake ..
make

Running

After building, you can run GenAlgo with the desired input image and options:

./genalgo -i input_image.png --period 100 --output result.svg

License

This project is licensed under the MIT License.