Location: Quality and Safety Assessment Research UnitTitle: Rapid identification of foodborne bacteria with hyperspectral microscope imaging and artificial intelligence classification algorithms
|KANG, RUI - Jiangsu Academy Agricultural Sciences|
|OUYANG, QIN - Jiangsu University|
|REN, NI - Jiangsu Academy Agricultural Sciences|
Submitted to: Food Control
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/24/2021
Publication Date: 6/25/2021
Citation: Kang, R., Park, B., Ouyang, Q., Ren, N. 2021. Rapid identification of foodborne bacteria with hyperspectral microscope imaging and artificial intelligence classification algorithms. Food Control. https://doi.org/10.1016/j.foodcont.2021.108379.
Interpretive Summary: A series of foodborne pathogen outbreaks often threaten the public. Five common pathogens including Campylobacter, Escherichia coli, Listeria, Salmonella, and Staphylococcus are responsible for major infections. The researchers from USDA have developed a hyperspectral microscope imaging (HMI) technique to classify these foodborne bacteria at the cellular level. However, the data analysis of HMI technology is still challenging. In this study, state-of-the-art artificial intelligent (AI) classifier named long-short term memory (LSTM) network was proposed for automated classification of foodborne bacteria. Furthermore, the spectra extracted from different regions of interest (ROI) of single bacterial cell were explored for improving classification accuracy. Compared to commonly used classification models using principal components analysis (PCA) techniques, our newly proposed AI classifier was able to identify unknown samples instantly with 92.9% accuracy using center ROI dataset from bacterial cells.
Technical Abstract: An artificial intelligent (AI) assisted hyperspectral microscope imaging (HMI) method was successfully developed to differentiate five common foodborne pathogens simultaneously. HMI was extremely powerful for characterizing living cells, with spectral information from every pixel of the cell. Three regions of interest (ROI) including whole-cell ROI, boundary ROI (outer membrane of cells), and center ROI (inner area of cells) were investigated to explore their performance for classification task. An artificial recurrent neural network named long-short term memory (LSTM) network was employed and optimized for directly processing the spectra acquired from different ROIs. Compared to classifiers based on principal component analysis (PCA) such as latent discriminant analysis (PCA-LDA, 66.0%), k-nearest network (PCA-KNN, 74.0%), and support vector machine (PCA-SVM, 85.0%), the AI-based classifier achieved the highest accuracy of 92.9% with center ROI dataset. Furthermore, AI-assisted HMI was able to identify foodborne pathogens instantly by eliminating complicated two-steps PCA feature analysis.