The Warsaw University of Life Sciences (SGGW) has developed a method that will allow wholesale milk buyers to quickly and effectively check whether raw milk delivered to a dairy is ‘fraudulent’. The authors of the new solution emphasise that its application will allow the elimination of so-called milk fraud already at the stage of supplying raw milk to dairies. 

Called MFA (Milk Fraud Analyzer), the solution is based on artificial intelligence and neural networks. It was developed by the team of Dr Marcin Golebiewski, director of the Institute of Animal Science.

“The developed method allowed to automatically check whether milk has been centrifuged to remove somatic cells from it before it is delivered to the dairy. This usually occurs when producers want to hide the fact that the milk comes from sick cows. In addition, this kind of illegal procedure is currently almost completely undetectable using standard analytical procedures”, says the head of the research team.

To detect milk fraud based on centrifugation, it is sufficient to look at the fat globules contained in the raw material. Although the test appears simple, so far milk consumers need the support of specialists and the right equipment, above all microscopes, to perform it. It is worth mentioning that, despite the use of a microscope and human assistance, examination by this method is subject to a certain degree of error. 

“The Milk Fraud Analyzer replaces human eyes and speeds up analysis. Testing consists of preparing and applying a milk sample for microscopic examination and taking a photograph of the sample. The photograph is transferred to our system, where it is automatically analysed. This is based on algorithms, artificial intelligence, and a neural network. The MFA immediately verifies the uploaded image and sends back feedback on whether the milk tested was centrifuged”, explains Prof. Golębiewski.

Arkadiusz Słomczyński