What is the status of visual recognition in the catering industry? 4 examples (+videos)

Why is there so much talk about visual recognition?

The challenges that this technology makes it possible to meet in catering are significant:

• Reduce rush hour queues
• Simplify the restaurant experience
• To save time and serenity
• Make the checkout process easier
• Accelerate collection
• Optimize the operational efficiency of a site
• Develop your knowledge of the consumer and refine your culinary offer...

Today, and even if their deployment is still in its infancy, several innovative solutions already exist on the market. This evolution is all the more important as consumers are becoming more and more accustomed to using self-service checkouts during their daily purchases (supermarkets, etc.).

1) Foodvisor and the Elior Group

In association with the French startup Foodvisor (a leading player in image recognition and calculation of nutritional information by artificial intelligence) via its subsidiary Trayvisorthe Elior group aims to deploy in its restaurants an innovative solution for visual recognition of meal trays at the checkout.

Here's how the solution works:

Based on algorithms using machine learning, the dishes on the tray are recognized at the checkout. The solution instantly identifies each dish or product, associates it with a price and immediately generates a paper or digital receipt via the TimeChef application.

"This technological innovation is a major contribution to the well-being of our guests: a simplified path through the restaurant means more time and serenity to fully appreciate the quality of meals. After several months of conclusive tests in a restaurant in the Paris region, we are now deploying this innovation in our restaurants since September 2019."explains Frédéric Galliath, General Manager of Elior France's corporate market.

2) Deepomatic and Compass Group's Express Terminal

Compass Group France has launched its Express Kiosk, a true "smart cashiering station". This self-service checkout integrates advanced visual recognition technology and allows customers to benefit from a checkout in just a few seconds.

After serving himself, the guest places his tray under the terminal equipped with a camera. In less than 300ms, the neural network of the solution allows him to identify the entire content of the tray. The consumer can then pay in less than 10 seconds, thanks to his badge.  

When in doubt about a product, the kiosk offers a short list of references so that the guest can select the corresponding item with one click. This approach is all the more interesting as it allows the terminal to learn from the guests' feedback, since each manual entry enriches its learning process.

The Express Kiosks were designed by the French start-up Deepomaticspecialized in image recognition. The latter has developed the different dimensions of the cash register:

• An easy-to-use cash terminal equipped with a camera
• Neural networks trained with Deepomatic Studio® software on a database of 1 million photos, able to recognize more than 10 000 products and/or recipes
• Deepomatic Run® software

‍

3 ) Aeye-Go and Sodexo

After meeting at the show, Sodexo Ventures and start-up Aeye-Go presented their solution for visual recognition of meal tray contents, coupled with a self-checkout system, at VivaTech 2019.

In China, it is even possible to use facial recognition to be identified and pay!  

Using high-definition cameras and image recognition technology, the solution automatically identifies everything on the tray and assigns it a name and price, as well as nutritional data. To achieve this, it took Aeye-go 2 years to build a database of nearly 100 million products.

Here again, the solution enhances the guest's experience, going so far as to allow them to do without a badge, smartphone or money when they go to the canteen... in a country other than ours for the moment, but the technology is there!

4) Livefood.ai by Belive.ai

Livefood.ai is a tray solution that automatically collects food:

• User recognition after identification
• Recognition of the products present on the tray
• Automatic collection
• Billing

This allows sites to absorb the influx of people during peak hours, to support employees in terms of health and well-being, to get closer to the core business, to better understand the guests and to optimize the performance of the checkout lines.

The solution does not require additional equipment, as its self-checkout technology is based solely on its cameras.

Visual recognition solutions are therefore already a key component of tomorrow's catering industry. At the same time, some of the staff who used to work at the cash registers can be trained to use the kiosks and to take on higher value-added service and reception tasks.

(c) main image: Elior Group