IFMG’s cheese-inspection app achieves 97% label-reading accuracy in tests

Source: br.edairynews.com
126 EN 中文 DE FR عربى
Researchers at the Instituto Federal de Minas Gerais (IFMG), Campus Bambuí, are developing artificial-intelligence tools to register and inspect Queijo Minas Artesanal. Project coordinator Calebe Giaculi Júnior said the inspection application achieved 97% accuracy in label-reading tests, including on crumpled or poor-quality labels.
IFMG’s cheese-inspection app achieves 97% label-reading accuracy in tests

The project, IA & Queijo Minas – Inovação na Tradição, uses computer vision to analyse photographs of artisanal cheese and automate production records and inspection tasks. Developed within the campus’s Centres of Excellence, it combines technology with knowledge of cheesemaking. Its proposed analysis covers four characteristics: maturation, rind type, geographical origin and producer identity. The system draws on photographs taken at cheese dairies, field applications and a database containing production codes, batch details, producer identification and dates.

Giaculi Júnior, an electrical-engineering lecturer and member of IFMG’s Department of Engineering and Computing, conceived the initiative after living in Serra da Canastra. His contact with local cheesemakers informed the project’s focus on production challenges and regional identity. “It was by living there, sharing daily life with producers, that I understood in practice the challenges faced by those who make cheese and the importance of protecting cheese as territory and identity,” he said. He is an electrical engineer with a master’s degree in electrical engineering.

Join the 2nd MENA Dairy Congress on January 24–25 in Al Ain, UAE — a key meeting point for dairy industry leaders, producers, investors, experts, and technology providers from across the MENA region and beyond.

Discover new market opportunities, exchange expertise, build valuable business connections, and discuss the future of the dairy industry in one of the region’s most dynamic markets.

Register here: https://menadc.dairynews.today

The team has developed a prototype that automatically registers production using computer vision and artificial intelligence. According to Giaculi Júnior, it has been validated and has led to a patent application that is being submitted. Two applications handle different stages of the process. The first, Aplicativo de Aquisição, lets producers photograph cheese at their own dairies; the software processes the picture and enters information into the database. The second, AuthProduto, is intended for inspection: it scans the cheese, reads its label through optical character recognition and immediately displays the result.

Researchers are collecting further photographs to extend validation. Their initial target is 5,000 images, followed by a collection of 150,000 to develop a more robust system. They intend to use the larger dataset to broaden testing across different production conditions. The team also plans to take trials beyond Serra da Canastra to other cheesemaking regions, including Serro, Araxá and Cerrado. Giaculi Júnior said the work remains under development.

A central objective is to improve traceability and address false geographical identification, in which cheese is presented as originating in a region that does not match its actual provenance. The proposed monitoring system would follow batches and identify potential irregularities. Giaculi Júnior said it could also help identify anomalies earlier, reduce waste and automate information used in certification procedures. The project aims to protect geographical origin and add value for producers who meet production and territorial-identification requirements. The team eventually wants the inspection application to support regulatory checks and help consumers verify authenticity.

Giaculi Júnior described participation in the International Forum on Family Agroindustry at ExpoQueijo Brasil 2026, in Araxá, as an important opportunity to meet the sector. Researchers demonstrated their existing tools and received comments from producers, associations and inspection authorities. “It was very important because it took the project out of the laboratory and brought us face to face with producers, associations and inspection bodies,” he said. Those contacts also opened discussions about partnerships to sustain the work and extend its application.

The project team comprises Giaculi Júnior; Fabiana Couto, a biologist with a doctorate at IFMG Campus Bambuí; computer engineer Marçal Moreira; and food engineer Alessandra Barcelos, who has postgraduate training in Data Analytics. People interested in learning about, supporting or investing in the initiative can contact Giaculi Júnior at calebe.giaculi@ifmg.edu.br or through WhatsApp on +55 34 98889 6420.


Key News of the Week
November 2026
  • Mo
  • Tu
  • We
  • Th
  • Fr
  • Sa
  • Su
  • 1
  • 2
  • 3
  • 4
  • 5
  • 6
  • 7
  • 8
  • 9
  • 10
  • 11
  • 12
  • 13
  • 14
  • 15
  • 16
  • 17
  • 18
  • 19
  • 20
  • 21
  • 22
  • 23
  • 24
  • 25
  • 26
  • 27
  • 28
  • 29
  • 30
Calendar