greg-kennedy
commented
6 years ago How fun, I like the "image processing" style entries. There was a similar entry to this from 2015, it was called The Cover of The Sun Also Rises.
The pictures at the end of your PDF really make this entry. I like being able to see the "color difference", it's really cool.
I was planning on making some awesome thing with AI agents and interactions etc. but it turns out I don't have the knowledge or time. Instead I made this, whatever this is. When given an image (or it can download a random one from lorempixel.com given a size) it makes a pdf of the closest color names it can find pixel by pixel. At the end it includes three images -- the original image, the image created using the RGB values of the colors, the difference between the two in red-scale and the difference multiplied by 10. The color names are all gathered from random websites (the attribution is in the text file). The image I used for the example is a totally free use image as far as I know.
(Picture.txt should really be picture.py and is my code, but I didn't want to make a whole repository just yet and apparently you can't attach .py files) picture.py colors.txt (this is the database of colors, it's terribly formatted so I just made the code able to handle it, sorry! If you want to add more it needs the name followed by something starting with a pound symbol, followed by an RGB formatted as (r,g,b) with no spaces in between)
Here's the output novel for a free use image I downloaded and scaled down. Scaled down it's a 260x195 image and the word count is 88050 words, not counting the title or image label, so keep that in mind if you make your own novel with it. out.pdf