Sesame oil is a popular condiment in China and other Asian countries, but its premium price has led to fraud by the sale of cheap vegetable oil adulterated with sesame oil essence.
One of the drivers for sesame oil’s popularity is its role as a flavour enhancer, but it is also viewed as a health ingredient.
Sesame oil is rich in fatty acids, sesamin, sesamolin, sesamol, tocopherols and inorganic elements, that have been linked to positive health benefits, and opened up lucrative export markets for Chinese producers in other regions of the world including the US and Europe.
Now, researchers from Yanshan University in China have developed a technique to speed up the testing of sesame oil samples based on a combination of 3D fluorescence spectroscopy and machine learning using AlexNet, a convolutional neural network or CNN.
3D fluorescence spectroscopy is already becoming a go-to technology for analysing oil samples, but is somewhat compromised because vegetable oils have limited fluorescence spectra, making differentiation of one type from another challenging.
Adding the machine learning element to the analysis has improved the accuracy – in fact, the researchers suggest that their technique was able to identify whether a sample was counterfeit with 100 per cent accuracy, whilst simultaneously predicting the concentration of sesame oil essence.
Importantly, the testing doesn’t destroy the sample, and the AlexNet CNN analysis “provides ideal and abundant data information for statistical analysis” with a short experiment time, as it sidesteps a “cumbersome tuning process and…dependence on high-performance computers.”
“This provides a new method for the field of food safety and quality identification, which is important for maintaining the normal operation of the oil market,” say the scientists. “Once the counterfeit sesame oil is sold in the market, the health of consumers and the normal operation of the market will be damaged considerably.”
The research is published in the journal Food Chemistry.
Credit: Google News
