Researchers have improved an AI tool designed for the rapid detection of bacterial contamination in food, significantly reducing misclassifications of food debris as bacteria. Traditional detection methods for contaminants like E. coli, Listeria monocytogenes, and Bacillus subtilis can take up to a week. In contrast, this new deep learning model can identify contamination within just three hours. Developed by Luyao Ma and collaborators from various universities, the model was trained to differentiate between live bacteria and microscopic food debris. The previous model misclassified debris as bacteria over 24% of the time, whereas the enhanced version virtually eliminated these misclassifications. The study, published in npj Science of Food, highlights the importance of early pathogen detection to prevent outbreaks and costly recalls. Researchers are now focused on optimizing this AI technology for broader applications within the food industry, ultimately improving consumer health and safety.
Share
Read more