System for Visual Recognition of Shelved Products - SOVA

Project Summary

System for Visual Recognition of Shelved Products (SOVA) is an IRI project coordinated by Cloudonia d.o.o. in collaboration with the Faculty of Electrical Engineering and Computing (partner). 

Image analysis of shelved products in supermarkets is important for optimal product management in the retail industry in order to minimize sales losses and increase the number of customers.

The project goal is to develop a methodology for the automatic analysis of images containing shelved products in supermarkets. The image analysis problems are solved using deep-learning methods.


Specific objectives are to develop:

  • methods for detection of shelves
  • methods for detection of price tags and recognition of their content
  • methods for product out-of-stock detection on shelves
  • methods for detection and classification of groups of shelved products
  • a cloud-based mobile application for product image analysis

Project Team

  • Prof. Sven Lončarić, principal investigator
  • Prof. Marko Subašić
  • Prof. Zoran Kalafatić
  • Lovre-Antonio Budimir, Ph.D. student
  • Branimir Filipović, Ph.D. student
  • Jeronim Matijević, master's degree student
  • Franko Šikić, Ph.D. student
  • Katarina Tolja, Ph.D. student

Project Duration

The project duration for the Faculty of Electrical Engineering and Computing is 24 months, starting from September 2021.




To be added


This research has been funded by the European Union through the European Regional Development Fund, under the grant KK. The project is co-funded by the company Cloudonia d.o.o and by the Faculty of Electrical Engineering and Computing.