/Pixit

A little project to create pixelated images.

Primary LanguagePythonMIT LicenseMIT

Pixit

It's downscaled but upscaled 😵 !

Demo

Contents

  1. Introduction
  2. How to run
  3. Logic behind pixelating
  4. Credits

Introduction

This is a little project which pixelate the given image.After each run program will show you resultant image and you can save it with desired name.

How to run

  1. Fulfill requirements.txt (i.e. pip install -r requirements.txt).
  2. Run with command python main.py <file_path>.
  3. Optional arguments
    • Intensity: python main.py <file_path> -size 20(Fill 20x20 pixels at once).
      Default value of size is 8.
    • Grey: python main.py <file_path> -grey(converts image to greyscale).
      By Default output images will be same as provided(in terms of channels).
    • Help: python main.py --help(To see help).

Some Examples

  1. python main.py "Test/test1.jpg"

    Original:
    Output:

  2. python main.py "Test/test2.jpg" -size 10 -grey

    Original:
    Output:

  3. python main.py "Test/test3.jpg" -size 100

    Original:
    Output:

  4. python main.py "Test/test4.jpg" -size 6 -grey

    Original:
    Output:

  5. python main.py "Test/test5.jpg" -size 20

    Original:
    Output:

Demo Console log

C:\Users\Deep Raval\Desktop\Projects\Pixit>python main.py "Test/test4.jpg" -size 6 -grey
Recieved Image with Shape: (1080, 1920, 3)
Converting Image to greyscale (if it is not already greyscale)...
Now shape of the Image is: (1080, 1920, 1)
Now Converting Image...
Save ? <y/n>: y
Enter Name for for file: test4o

Logic behind pixelating

As we know images are simply array of pixels. For ex. a RGB image of resolution 1333x688 can be viewed as an array of shape (1366, 768, 3)(last 3 for seprate Red, green and blue pixels).Now let's take SIZE = 10 and pixelate the image of same size.
First let's start with first pixel (i.e. (0, 0, 0-3)) and give square with 10x10 starting from selected pixel same pixel value.Now next we must select pixel which is atleast SIZE distance away from current pixel(i.e (11, 0, 0-3) in this case). After repeating this process for whole image we can pixelate an image. Note that if image's resolution is not perfectly divisible by SIZE then there maybe some pixels unchanged.
I hope you understood.
P.S.: I know it's better to give square mean value of pixel.But, hey this also looks good.xD

Credits

Contributors 💻 :

Without these excellant libraries ❤️ this would not have been possible.

  • opencv-python
  • numpy
  • matplotlib