CANBERRA, Aug. 12 (Xinhua) -- Research led by Australian scientists has shown that artificial intelligence (AI) can detect contaminated food before it reaches consumers, potentially preventing millions of deaths annually.
Researchers have shown that hyperspectral imaging (HSI) combined with machine learning can quickly and accurately identify mycotoxins -- toxic compounds produced by fungi -- in cereal grains and nuts during growth, harvest and storage, according to a statement released Tuesday by the University of South Australia (UniSA).
Mycotoxins are linked to cancer, compromised immunity and hormone disorders. Foodborne contamination, including mycotoxins, causes 600 million illnesses and 4.2 million deaths worldwide each year, according to the World Health Organization.
Traditional detection methods are slow, costly and destructive, making them impractical for large-scale real-time food processing, said UniSA PhD candidate Ahasan Kabir, lead author of the study, published in the Switzerland-based journal Toxins.
"In contrast, hyperspectral imaging -- a technique that captures images with detailed spectral information -- allows us to quickly detect and quantify contamination across entire food samples without destroying them," Kabir said.
HSI uses spectral "footprints" to quickly identify contaminated grains and nuts when paired with machine learning based on subtle spectral variations, said Kabir, collaborating with researchers from Canada and India.
The study reviewed more than 80 recent trials and found that machine learning-integrated HSI consistently outperformed conventional testing, particularly for detecting aflatoxin B1, one of the most dangerous foodborne carcinogens.
The technology could be deployed on processing lines or used in handheld devices to inspect products such as wheat, maize, almonds and peanuts, reducing health risks and trade losses, said the authors, who are currently refining the technique to improve accuracy and reliability. ■
