Task
Cut queues in canteens and food courts at peak hours and reduce dependence on cashiers.
The menu changes daily, so the checkout has to learn new dishes quickly and without developers.
foodscanner.tech
~10 min
Foodscanner — an AI checkout for canteens
To learn a new menu.
Computer visionEdge inferenceSelf-checkouts
~10 min
to learn a new menu
~98%
recognition accuracy
An AI self-checkout for canteens. A new menu of 150+ items is learned in about 10 minutes.
Client
foodscanner.tech
Industry
Foodtech, corporate catering
Areas
Computer vision, edge inference
ESTESIS role
Own product
From task to result
Client
Foodscanner is ESTESIS's own product for corporate canteens and food chains. Its deployments include the canteen at Sber headquarters.
What we did and how
We developed the whole product: recognition models, the checkout software and the process of learning a new menu.
- 01Recognition of dishes on a tray from the camera image
- 02Learning a new menu of 150+ items in about 10 minutes
- 03Receipt generation and integration with the checkout software
- 04Running on the device without depending on the cloud
Result
~10 min
to learn a new menu
~98%
recognition accuracy
The checkout recognizes dishes with about 98% accuracy and learns a new menu in about 10 minutes.
Stack and competencies
More stories
Your task —next story
Tell us about the project — within 5 working days we will send an architecture and an estimate.
Accredited IT company · Russian Ministry of Digital Development registry No. 20142