AI Generates Images From Text-based Recipes

Artificial intelligence (AI) researchers are developing tools that can generate images from recipes and vice-versa, helping consumers create their own dishes at home and making cooking more intuitive and accessible for all.
Computer scientists at Tel Aviv University in Israel have built an AI algorithm that can conjure an image of any dish using just a text-based recipe listing key ingredients and outlining a specified method of preparation.
It achieves this by analysing and converting text into numerical values and running these numbers against descriptors of more than 50,000 photos of food from within in its existing image database via a process called text embedding.
Tackling the reverse of this process, researchers from Facebook's tech innovation branch Fair have developed AI software that can determine the recipe used to make any dish simply by analysing a photo.
Joelle Pineau, head of Fair's Montreal lab, told Forbes: "Everyone is always taking pictures of their meals these days. Sometimes there's ingredients you can see, but there's also ingredients you can't always see, like sugar and salt and things like that. So, they train the AI with pairs of images and recipes. But then when they test it, they just give the image and it generates a recipe."
This technology has great potential for foodie consumers, who could soon be able to eat out at a restaurant, and then recreate the dish in the comfort of their own home.
See Activating At-Home Foodies for more on smart kitchen tech assisting budding home cooks to create exciting dishes, and read Kitchen of the Future for a wider understanding of new tech-infused cooking habits and rituals.