Researchers at MIT say they have developed a piece of software that can analyse the picture of food and bring back the recipe required to make it.
Called Pic2Recipe,MIT says it currently has a 65% success rate in identifying a recipe from a simple picture of food.
To make the service work, a team from MIT’s Computer Science and Artificial Intelligence Laboratory created a database of around 1 million food recipes - this needs to be big because the software doesn’t actually look at a picture and guess the recipe, instead it lists all the ingredients it thinks are in the food - this data is then referenced with the database and users are sent 10 recipe suggestions that best match the picture recieved.
The demo page is currently not online, which makes it hard to test the claims, but the video below shows what you can expect from the service.
NFTs have taken the digital realm by storm, with many of the crypto-assets being sold for astronomical fees. But how can restaurants and food professionals explore the possibilities of this new technology? FDL takes a look.