dmartzol / poly

Optimizing a fixed number of polygons to fit an image

Poly

Go CLI application to optimize the shape of a fixed number of polygons to minimize the difference to a target image using a genetic algorithm just for fun.

Project highly inspired in the work of Michael Fogleman and Roger Johansson but using a different approach and getting different results.

Target image

Resulting image

Image generated with 200 polygons and 80000 iterations.

Features

• Select the number of polygons used to reproduce the target image.
• Select the number of generations for the algorithm to iterate.

Installation

Install with:

go get -u https://github.com/dmartzol/poly.git

Usage and examples

Options

Usage: polygonal [OPTIONS] -o output
  -i string
    	input image path
  -n int
    	number of iterations (default 1000)
  -o value
    	output image path
  -p int
    	number of polygons (default 50)
  -r int
    	resize large input images to this size (default 256)

To generate an image with 200 polygons and 50k iterations input:

poly -i input.png -o output.svg -n 50000 -p 200

TO DO

• Concurrency
• Implement save a frame every N iterations.
• Add more examples.
• Improve this README.

License

This project is licensed under the MIT License

Examples