Deep learning and multiwavelength fluorescence imaging for cleanliness assessment and disinfection in food services

Authors: GORJI, HAMED - University Of North Dakota; Van Kessel, Jo Ann; Haley, Bradd; HUSARIK, KAYLEE - University Of North Dakota; Sonnier, Jakeitha - Jackie; SHAHABI, SEYED - University Of North Dakota; ZADEH, HOSSEIN - Collaborator; Chan, Diane; Qin, Jianwei - Tony Qin; BAEK, INSUCK - Orise Fellow; Kim, Moon; AKHBARDEH, ALIREZA - Collaborator; SOHRABI, MONA - Collaborator; KERGE, BRICK - Collaborator; MCKINNON, NICHOLAS - Collaborator; VASEFI, FARTASH - Collaborator; TAVAKOLIAN, KOUHYAR - University Of North Dakota

Submitted to: Frontiers in Remote Sensing
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 8/22/2022
Publication Date: 9/22/2022
Citation: Gorji, H., Van Kessel, J.S., Haley, B.J., Husarik, K., Sonnier, J.L., Shahabi, S., Zadeh, H., Chan, D.E., Qin, J., Baek, I., Kim, M.S., Akhbardeh, A., Sohrabi, M., Kerge, B., Mckinnon, N., Vasefi, F., Tavakolian, K. 2022. Deep learning and multiwavelength fluorescence imaging for cleanliness assessment and disinfection in food services. Frontiers in Remote Sensing. https://doi.org/10.3389/fsens.2022.977770.
DOI: https://doi.org/10.3389/fsens.2022.977770

Interpretive Summary: Foodborne illness is a global concern and there are about 48 million cases in the US annually. Contamination of food products can occur from farm to fork. The food service industry requires precise, reliable, and speedy contamination detection and disinfection methods to ensure a safe product is delivered to consumers. Contamination in food-related services can cause foodborne illness, endangering customers and jeopardizing provider reputations. Fluorescence imaging has been shown to be capable of identifying organic residues such as food and biofilms that can host harmful microorganisms. We used a new fluorescence imaging technology, applying state-of-the-art deep learning algorithms to identify and segment contaminated areas in images that were taken of equipment and surfaces in commercial food preparation areas. Deep learning models demonstrated a 98.78% accuracy for identifying contamination on various surfaces. The portable imaging system also has a built-in disinfection capability that was evaluated on three groups of bacterial pathogens (S. enterica, E. coli, and L. monocytogenes). Results showed that the system was capable of significantly reducing bacterial contamination in under 5 seconds. The results of this study determined that these new technologies could be used by the food service industry to help provide safe, uncontaminated foods to consumers.

Technical Abstract: Precise, reliable, and speedy contamination detection and disinfection is an ongoing challenge for the food-service industry. Contamination in food-related services can cause foodborne illness, endangering customers and jeopardizing provider reputations. Fluorescence imaging has been shown to be capable of identifying organic residues and biofilms that can host pathogens. We use new fluorescence imaging technology, applying state-of-the-art deep learning algorithms to identify and segment contaminated areas in images of equipment and surfaces. Deep learning models demonstrated a 98.78% accuracy for differentiation between clean and contaminated frames on various surfaces and resulted in an intersection over union (IoU) score of 95.13% for the segmentation of contamination. The portable imaging system's intrinsic disinfection capability was evaluated on S. enterica, E. coli, and L. monocytogenes, resulting in up to 8-log reductions in under 5 seconds. Results showed that fluorescence imaging with deep learning identification algorithms can help assure safety and cleanliness in the food-service industry.