Selecting a retail outlet on an interactive map, building routes, filtering stores
Merging several photos of a long shelf and checking the quality of photos
Automatic comparison of product display against planogram in real time
Identification of discrepancies on the shelves and sending a report to the manager immediately
Offline or online mode
Choosing a category for a photo (shelf, price tags, POSM)
Respond to problems while at the point of sale
Fraud protection
Image recognition is the assignment of an object to a certain class by highlighting the essential features that characterize this object. Objects are different things. Recognition of goods on the shelf is also image recognition.
First, experts prepare a training set. These are photos (renders) of the goods provided by the customer. Based on this set, the neural network is trained: it learns to recognize objects and correlate them with certain products. Next, the results are validated on real-world photos.
Product recognition in merchandising works as follows: users take photos of products on the shelf. Next, the photos are uploaded to the data center, where artificial intelligence analyzes and classifies them. Neural networks recognize images, that is, products, and determine whether they are displayed correctly on the shelves. Also, in real time, you can calculate various KPIs and identify out-of-stock.
One photo is processed in up to 30 seconds.
Product recognition accuracy is >95%.
The mobile application allows taking a photo offline. When the Internet connection is established, the photo will be uploaded for further processing.
Merchandiser can take a photo of several parts of the shelf in the application, and Goods Checker will merge them into one photo.
