Scientists at the Nanyang Technological University in Singapore in cooperation with colleagues at Monash University in Australia and Jiangnan University in China have developed a system to detect spoiled meat and poultry*.
The principle involves placing a barcode strip incorporating twenty porous nanocomposites of chitosan and dyes on a cellulose acetate base in inner packaging of the product. The strip reacts with volatiles produced by decaying meat. The status of the contents of the package can be ascertained within 30 seconds using a cell-phone reader incorporating ‘deep convolutional neural network (artificial intelligence) technology’.
The developers claim an accuracy of 98.5 percent in detecting spoiled meat products compared to 61.7 percent for current artificial scent-screening systems. The barcode and cell phone combination was tested with beef, fish and chicken stored at 25C. Accuracy was determined from measurements taken from six packages at different time intervals over 48 hours yielding 4,000 images. Spoiled meats could be identified with total accuracy and with a range of 96-99 percent sensitivity for either fresh or slightly spoiled meat products.
It is evident that the system, if applied commercially, would enable institutional kitchens and consumers to correlate freshness with ‘use by’ dates. Unnecessary disposal would be prevented contributing to sustainability.
*Guo, L. et al Portable food-freshness prediction platform based on colorimetric barcode combinatorics and deep convolutional neural networks Advanced Materials.doi.org/10.1002/8dma.202004805. October 1, 2020