Create a target image from given shapes like triangles, cubes, circles, ... . There are two kinds of images that work best, Highly Detailed Images: These images have lots of fine details and intricate features. When blurred, the details still hold enough information to create a recognizable image and Very Simple Images: These images have clear, distinct shapes and minimal elements. Even when blurred, the simplicity makes them easy to recreate. The worst image you can give it is one where color changes are not small enough to be ignored but still too complex, making it difficult to use the basic shapes effectively.
Arguments are the target image path, the shapes directory path, changing the color of the shapes, using random shapes, and finally if it should start at another image. (This doesn't work sometimes)
python3 shaper.py from_image.png ./shapes True False middle_image.pngYou can leave out the last parameter if you want to start fresh like this:
python3 shaper.py from_image.png ./shapes True FalseWhen you close the window it'll save the image to final_image.png
[On bigger images, otherwise it can lead to an opencv error]
- Python 3.12.1; Windows 11; Qt6
- Python 3.10.12; Ubuntu 22.04 LTS; Qt6 & libxcb-cursor0