Multi-label Classification

I'm submitting my Optical Illusions Dataseet paper to INSAM journal - horray! However, they have a 3000 word requirement, and the body of my paper is only 2200 or 2000 if you don't count the appendix. Adding more words is actually pretty easy, since I original wrote my paper in a terse writeup format. There are plenty of things I can expand on, using the classic achademic paper tropes: explaining basic concepts, including summaries of other papers, explaining the contents of a figure in words, making long-winded subjective analysis of results. These are all things I would rather not do, but they are fair game for increasing the word count to the required amout. In a perfect world my paper could be let to live as the scrawny little thing that it is, but I'll have to force feed it butter until it weights enough for market.

I'm also (perhaps foolishly) adding an additional experiment. I'm going to do multi-label classification, which is what the MoIllusions dataset should have had done to it in the first place.

This is going to be a little bit tricky, but it's mostly a data task. I have the directory of images, and JSON for the labels. The workflow will look like this:

This tutorial is going to be my base for how to work with images: