Polyvolve

A genetic algorithm that recreates images using transparent polygons. See http://ericdavidwells.com/projects/polyvolve for examples and details

Dependencies

• Python 3.7.2
• OpenCV 4.1.1.26
• Matplotlib 3.1.2
• Scikit-image 0.16.2

Getting Started

To see the program in action, clone the repository, set up the dependencies, and run python main.py -d -p.

Command Line Options

-h, --help\

• show this help message and exit

--image_path IMAGE_PATH\

• Path to source image.

--output_dir OUTPUT_DIR\

• Where output images should be stored.

-x X\

• Output image width.

-y Y

• Height of output image. If not set, automatically calculates using the input image's ratio.

--display_x DISPLAY_X

• Width of displayed image.

--display_y DISPLAY_Y\

• Height of displayed image. If not set, automatically calculates using the input image's ratio.\

--calc_x CALC_X
--calc_y CALC_Y\

--num_individuals NUM_INDIVIDUALS\

• Population of individuals (aka images) present at each time step.

--num_polygons NUM_POLYGONS\

• Number of polygons per individual

--num_vertices NUM_VERTICES\

• Number of vertices per polygon

--mutation_rate MUTATION_RATE\

• Rate at which individuals mutate.

--mutation_amount MUTATION_AMOUNT\

• The degree to which an individual is mutated.

--p_survival P_SURVIVAL\

• Percentage of individuals that survive to the next step.

--p_reproduce P_REPRODUCE
Percentage of individuals which reproduce each step.

--init_fn {random,zero,zero_color,random_rect,random_circle}\

--fitness_fn {ssim,mse,mseLAB}\

• Function for determining the fitness of a given individual.

--crossover_fn {random,two_point}\

• Function for determining how individuals cross over (eg. what their children look like.)

--mutation_fn {random_shift}

• Function for mutating an individual.

-s

• If selected, will periodically save the best individual, as an image, to diskm as well as a graph of fitness over time.

-p\

• If selected, will display a plot of the program fitness as the program executes.

-d

• If selected, will periodically display the best individual.\

--minimize_fitness

• If selected, chooses individuals with minimum fitness.

--max_generations MAX_GENERATIONS\

• Maximum number of generations before the program finishes execution.

--generations_per_display GENERATIONS_PER_DISPLAY\

• Number of generations between each image being displayed.

--generations_per_plot GENERATIONS_PER_PLOT\

• Number of generations between each plot update.

--generations_per_save GENERATIONS_PER_SAVE\

• Number of generations between saving images.

--run_to_fitness RUN_TO_FITNESS\

• Run until a fitness of the given value is reached (should be from 0 to 1).

--save_last

• Saves final frame.

Image Source

starrynight.jpg was downloaded from https://en.wikipedia.org/wiki/File:Van_Gogh_-_Starry_Night_-_Google_Art_Project.jpg.