PixelSort

A Python pixel sorting utility created using Pillow (a fork of the Python Image Library). Use it to create unique artistic effects with your favourite photos!

Install

In order to use the script, first install Pillow:

pip install Pillow

Then edit the parameters at the beginning of the script to generate new images.

Example

Using the parameters:

 mode = SortMode('row', 'l', 'brighter', 'sort', -1, 250, 20, 150, 40)

And setting the file path:

im = Image.open("./Example Images/Mixed Signals.jpg").convert("HSV")

will generate the same example image from above!

Usage

All script parameters are controlled by editing the first line of main. From left to right they are:

• Row/Column Specifier: 'row' for sorting rows, 'col' for sorting columns.
• HSL Selector: choose how you want to sort each pixel: by Hue (color), Saturation (intensity) or Luminance (brightness))
• Range Selection Mode: 'brighter' for selecting pixel ranges with a higher luminance than the set black point, 'darker' for selecting pixel ranges with a lower luminance than the set white point
• Sort Mode: 'sort' for sorting pixels by their HSL values, 'average' for replacing each selected range with the average pixel from that range (plus the luminance bias factor)
• Direction: 1 for left-to-right/top-to-bottom, -1 for right-to-left/bottom-to-top
• Noise Threshold: sets the minimum size in pixels a range must be in order to be sorted
• Black Point: sets the black point for 'brighter' range selection
• White Point: sets the white point for 'darker' range selection
• Luminance Bias: set the luminance increase for pixels when using the 'average' mode

Efficiency Improvements

Pixels are sorted using a bitwise base 16 implementation of the Radix-LSD sorting algorithm in order to improve run-times on large images.

Rows/Columns are processed on separate cores simultaneously through use of the Python multiprocessing library. In order to maximize performance, change the processes parameter on line 46

pool = mp.Pool(processes=4)

to equal the number of cores availible on your local machine.

License

This project is licensed under the terms of the MIT license.