Food Recognition models training for autonomous checkout

When lunch time comes, canteens and restaurants tend to be full and many times, customers have to queue and pay before they eat. To avoid wasting time on the queue and process the orders faster, the solution is called autonomous checkout. Thanks to a food recognition algorithm involving computer vision, food can be recognized on the tray and the customer is automatically charged for what he has on his tray.

What part does isahit play?

For the algorithm to be able to recognize food on tray meals, it needs to be trained to do so, this is when isahit steps in. Our job involves annotating datasets of plenty of food images so the algorithm can tell and differentiate what it sees on the tray meals. It is therefore very important to annotate the datasets as better as possible so the precision of recognition is close to 100%. Thanks to our diverse workforce, we are able to leave out any kind of bias when it comes to annotate. Depending on the project, the language and the country of the client, we assign contributors accordingly in order to reduce bias as much as possible.