Researchers use AI to identify where EVOO came from

Researchers use AI to identify where EVOO came from

A new way to determine the authenticity of extra virgin olive oil has been devised by a group of researchers in Italy.

Their study, published by Food Chemistry, details a method that includes training artificial intelligence to identify the provenance of an extra virgin olive oil using its phenolic compounds and sterols.

The researchers used Taggiasca Ligure extra virgin olive oil from Liguria in northwestern Italy.

Yet the methodology we have deployed could be applied to any other extra virgin olive oil, any cultivar, in any region,” Luigi Lucini, researcher in the Department of Sustainable Food Processes from Università Cattolica del Sacro Cuore and co-author of the study, Olive Oil Times said.

See also:Using Isotope Fingerprints to Authenticate Olive Oil and Fight Fraud

The main driver of development has been the spread of the Taggiasca olive, native to the region, to other countries. Therefore, researchers felt it was important to be able to identify and label Taggiasca monovarietals from Liguria.

I hear that the Taggiasca cultivar is planted overseas, in places like Greece,” Lucini said. When we talk about wine, we are used to the notion of terroir. However, the link between extra virgin olive oil and the territory of origin is a reality and involves specific quality characteristics.

In the study, researchers said they could correctly identify locally produced Taggiasca Ligure olive oils 100% of the time.

We worked for four years on the project, and the last year was entirely devoted to training the system and verifying the effectiveness of the method,” said Lucini.

The research team compared the new method to the FaceID authentication tool widely adopted by smartphone manufacturers.

This system learns to recognize different angles of a specific face to allow access to the device,” Lucini said. Our method does the same thing; instead of somatic parameters, it recognizes chemical parameters, which allows it to authenticate the origin of the product.

Researchers began the project by creating a robust dataset using 408 samples of Taggiasca Ligure extra virgin olive oil collected over three harvest seasons. With the cooperation of local producer associations, they labeled each sample with contact details.

Using metabolomics, the chemical fingerprint of a specific cellular process, the researchers were also able to identify thousands of different compounds, dozens of which are unique to locally produced Taggiasca Ligure olive oil.

Cholesterol derivatives and phenolic compounds (tyrosols, oleuropeins, stilbenes, lignans, phenolic acids and flavonoids) were the best markers, based on statistics,” the researchers wrote. Our results reinforce the concept of terroir’ for extra virgin olive oil and indicate that profiling of sterols and phenolic compounds can support the integrity of extra virgin olive oil if adequate data treatments are adopted.

The extra virgin olive oil content varies from season to season,” Lucini added. Especially in Liguria, where you can find olive trees growing at sea level and others growing hundreds of meters only a few kilometers away.

Differences can also come from weather or farming techniques,” he said. That’s why we collected data in different seasons to determine the exact markers we needed.

See also:European geographical indicators valued at more than 80 billion dollars

Once the dataset was built, artificial neural networks were trained to identify Taggiasca Ligure extra virgin olive oils and deployed to determine the authenticity of oils labeled as such.

The researchers said the dataset should be flexible enough to determine whether locally produced blends, including those claiming to include the Taggiasca olive, really do.

Much like the FaceID tool, which recognizes me even if I’m wearing glasses, our method does the same, and it’s not obsolete when faced with olive oil blends,” Lucini said. With FaceD, this happens because wearing glasses is not a determining parameter. The same goes for our method.

To test the system, the researchers tested samples of blended olive oil, where non-Taggiasca oils made up 5 to 60 percent of the blend. Frantoio was one of several olive oils used in blends.

The reason we chose the Frantoio cultivar is the close genetic similarity to the Taggiasca cultivar, which is derived from the same olive tree as Frantoio,” Lucini said. This is due to monks adopting olive trees in the Middle Ages, with cultivars evolving from there.

Both cultivars have a common ancestor, and if you use common genetic analyses, the two cultivars are virtually indistinguishable,” he added.

However, the researchers have concluded that their artificial intelligence is ready to identify Taggiasca Ligure extra virgin olive oils outside of laboratories. Now, the research team’s next steps will focus on wine.

The reason is that we try to work on high value-added products that can justify such demanding work,” Lucini said.

For this reason, the new method is mainly reserved for products with Protected Geographical Indication (PGI) or Protected Designation of Origin (PDO) certification from the European Union.

Italy has 49 extra virgin olive oils with PDO or PGI certification, with several more applicants seeking their own geographic indicators in the coming years.

Extra virgin olive oil is one of the foods most exposed to fraud,” said Marco Trevisan, research coordinator and professor of food chemistry at Università Cattolica del Sacro Cuore.

And this is even more true for protected products, such as Taggiasca Ligure, for which consumers are willing to spend more money,” he concluded. Our work… is a relevant step for the protection of PDOs.

#Researchers #identify #EVOO

Leave a Comment

Your email address will not be published. Required fields are marked *