Food service & QSR
Kitchens, counters and food waste, labelled dish by dish.
Food recognition, kitchen and drive through video, order accuracy and food waste annotation for restaurant and food AI.
From the kitchen to the bin
Every part of the workflow where we can take work off your team.
Kitchen
Counter and drive through
Food waste
Menu and data
We work with kitchen video, counter cameras, food images and drive through video.
What we label
Every class is written into a guide with your team before the first batch. Files come back as COCO JSON, YOLO, Pascal VOC, CSV, JSONL or your own schema.
Food images
Dishes and ingredients
Class and polygon
Burger, fries, rice, salad, sauce
Portion and state
Class and attributes
Full, half, eaten, spoiled, burnt
Kitchen and counter video
Prep steps
Time span and class
Assemble, cook, pack, hand over
Orders
Bounding box and class
Item present, item missing, wrong item
Queues
Bounding box and track
Car, person, waiting, served
How we test the model
The tests that matter for food service & QSR models, and what you get back from each.
Recognition accuracy
Model labels checked dish by dish against reviewed images.
You get backAccuracy by dish and lighting
Order accuracy checks
Detected bag contents compared with the order ticket.
You get backMissed and wrong items by station
Waste log accuracy
Logged waste compared with what reviewers see in the bin.
You get backError by food type and site
Where food service & QSR teams start
Problems teams bring us, and the first piece of work we would scope.
New menu items break recognition every season.
What we doNew dishes collected and labelled before launch, with look alike items logged.AnnotateWaste logs are guessed at the end of a shift.
What we doBin camera images labelled and checked live, as we do for Client D.Live reviewWrong orders leave the counter.
What we doBag and tray images labelled against the ticket, so the model learns what a complete order looks like.AnnotateRelated work
Twelve times the volume in five weeks, with quality holding at 95 to 98%
Human in the loop support for visual deviation detection on production lines, with a response time requirement under two minutes.
24/7 live review for production line visual AI, quality above 95% by month two
2,000 robot videos described every week by a team built in seven days
Half a million images for autonomous navigation, with every file kept on EU servers
Bring us the food service & QSR data your model gets wrong.
Send a small batch and get it back done, free, usually within 24 to 48 hours. You judge the work before you pay for anything.