Solutions based on machine learning
Image Alt

Label recognition

System for automatic monitoring of prices and reporting on the products presented

Price tag photorecognition service

Based on the data processing system, we have implemented a solution for recognizing price tag photos. Images are sent to the input, after processing, the data is obtained:

  • Name of product;
  • Price;
  • Promotional product Yes / No.

Depending on the requirements and type of labels, the list of fields for recognition is customizable, data such as can be added:

  • date;
  • second price;
  • barcode and others.

We are able to recognize the price tags of all major retail chains: Pyaterochka, Magnit, Dixie, Lenta, Auchan, METRO and many others.

With the help of our service, we have already identified more than 1,000,000 price tags in the food and fmcg segment as part of our project to create a unified price base for Moscow products and retail audit.

Algorithm of the program

For the detection of Areas, we use neural networks (deep learning).
Next, inside the area, we recognize the text using OCR (Optical Character Recognition).

1. We find the algorithm in the photo of the price tag area

2. Finds the area inside the labels.

3. Defines within the price tags the area Prices: rubles and kopecks

What is included in the basic version?

Recognition of one price tag in one photo

Set of fields: name / price / stock availability

Время распознавания одного ценника до 24 часов

Accuracy - 97%

Integration FTP, Email

What is included in the extended version?

Recognition of multiple price tags on one photo

The ability to add additional fields to the report: country, manufacturer, weight of goods, type of packaging, etc.

Possibility of integration with Personal Account, mobile application

Provision of EAS product codes

Creating a customer profile in a web service

Additional services

Classification of goods
Often on price tags, the name of the product is written in an abbreviated form, which complicates its further classification.
If necessary, we can determine the category, type of product, brand, type or taste of the product, unit of measurement, etc.
EAC code affixing
To reconcile data from price tags and the existing database of goods, for example, for inventory or price monitoring, it may be necessary to have a unique product code. We are able to determine the product EAC code by name, which allows us to unambiguously link the price tag to the database.

Trusted by clients from all over the world


Let’s discuss how machine learning
can be used in your area