So as we all know, Japan has 1 vending machine for every 23 people. That means you will have tons of different combinations of drinks available.
What I want to do is train some AI to recognize the drinks in photos of the vending machine (example of image in attachments).
All you have to do is do the recognition part, and label them at which position they are on the drink (left to right, top to bottom, 1-20/30, etc). So item 1 would be the first one in the top left, you can call it "Drink A" and I'll update the label later with the correct Japanese language/English.
I prefer this to be built in Python, since it will be used on a website's framework