Automatic analysis of the contents of store shelves, support of shelf audits and verification of product promotion and positioning rules.

The system enables fast, accurate and automatic analysis of the contents of store shelves, supporting shelf audits and verifying compliance with the rules of product promotion and positioning.

‍

‍

The analysis is based on photos taken with smartphones or tablets. Thanks to the use of advanced computer vision and machine learning algorithms, store products are automatically identified, counted and their mutual placement is verified. The results of the analysis are presented in the form of a report.

‍

‍

Basic parameters of the system:

‍

• Minimum photo resolution 800x600 pixels

• The average percentage of correct product recognition is 95%

• It takes about 1 second to analyze a single photo

• It takes about 48 hours to add the ability to recognize a new product

‍

‍

The flexible implementation of the system allows it to be deployed as a service on an external computing server. Thanks to the availability of the REST API, it is possible to use a dedicated graphical user interface as well as programmatic integration with already existing systems.

Detection of fronts of selected products, including location indication:

‍
The system can detect specific packages or entire product groups. The use of deep learning techniques makes it possible to teach the applied neural network and successively add new products to the pool of objects recognized by the system. Thanks to this, it is possible to flexibly extend the system and carry out full verification of the planogram and the study of shelf shares.

Depth Clogging Check:

‍
If the picture shows products positioned deep into the shelf, the system can identify these areas and count the packaging there. Based on this data, it automatically determines whether the number of products on the shelf complies with the previously established rules of availability.

Verification of the number and arrangement of products relative to each other:

‍
Information on the number and location of products is used by the system to check whether certain requirements and rules for product promotion and positioning are met. The system verifies the correct exposure of the goods, the number of visible fronts and their arrangement relative to each other, as well as the presence of packaging variants.

‍

‍