Have you ever come across a particular dish and wanted to know the recipe behind it? All of us have been there. And if you have watched HBO’s Silicon Valley, you would all the remember the hot dog and not hot dog app. The Researchers from the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL) have now come up with a neural network that can take a look at a specific dish and come up with the recipe for that particular dish. You can read the research paper over here.
Our Instagram feeds and Facebook walls are flooded with various food images and recipes daily. Now, a tool like this may also be applied to the enormous of food images shared on social media to get an insight into the significance of food and its preparation.
The researchers have built a database of over 1 million recipes that were annotated with information about the ingredients, and the data was used to train a neural network to find patterns and make connections between pictures of food images and the recipe. Over a period of time, the device was able to make refined connections to the various ingredients that would go into a dish.
Though it is not a finished product, the team has released an online trial. It was noted that this trial product worked the best on desserts and didn’t yield the desired results on sushi and smoothies.
This research paper will be presented at the Computer Vision and Pattern Recognition conference in Honolulu. This could also be adapted to provide nutritional information, MIT said.