|SEO, YOUNG-WOOK - US Department Of Agriculture (USDA)|
|Hinton, Jr, Arthur|
Submitted to: Transactions of the ASABE
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 12/1/2014
Publication Date: 2/20/2015
Citation: Park, B., Seo, Y., Yoon, S.C., Hinton Jr, A., Windham, W.R., Lawrence, K.C. 2015. Hyperspectral microscope imaging methods to classify gram-positive and gram-negative foodborne pathogenic bacteria. Transactions of the ASABE. 58(1):5-16.
Interpretive Summary: Food safety is worldwide issue for public health every year. There have been numerous outbreaks of foodborne illness during the past decade. With regard to serious outbreaks, Salmonella was responsible for most infections and incidents. Current detection methods for foodborne pathogens are limited for practical use due to the time-consuming nature, cumbersome results and sensitivity. Therefore, more sensitive, accurate and rapid pathogen detection method is needed for food industry. Hyperspectral imaging method is a candidate to meet the requirement for real-time, in-situ foodborne pathogen detection. Among several hyperspectral imaging platforms, acouto-optic tunable filter (AOTF)-based hyperspectral imaging system was developed for foodborne bacteria detection. The objective of this research is to develop micro-level optical methods to identify foodborne pathogens with AOTF-based hyperspectral microscope imaging technique. Specifically, classification models to identify gram-positive and gram-negative foodborne pathogenic bacteria were developed for rapid detection of foodborne pathogens from microcolony samples.
Technical Abstract: An acousto-optic tunable filter-based hyperspectral microscope imaging method has potential for identification of foodborne pathogenic bacteria from microcolony rapidly with a single cell level. We have successfully developed the method to acquire quality hyperspectral microscopic images from various gram-negative and gram-positive bacteria live cells. Among the contiguous spectral images from visible/NIR between 450 and 800 nm, the scattering intensity of spectral images were distinct at mostly visible wavelengths. Specifically, the scattering peak intensity was distinct at 458, 498, 522, 546, 574, 590, 646, 670 and 690 nm for Staphylococcus. Similarly, the distinct peak spectra were observed at 462,498, 522, 546, 574, 598, 642, 670 and 690 nm for Salmonella. For both cases, the scattering intensity of outer cell membrane was brighter than inner cell membrane except at 546 nm, which is possibly caused by excitation of metal-halide lighting source. The scattering intensity from a single cell varied with the wavelengths as well as the type of bacteria. The overall variability of intensity was 31.2% for gram-negative (Salmonella) and 42.7% for gram-positive (Staphylococcus) bacteria. With scattering intensity data from both five serotypes (Kentucky, Enteritidis, Typhimurium, Infantis, and Heidelberg) of Salmonella and five species (aureus, haemolyticus, hyicus, simulans, and sciuri) of Staphylococcus bacteria cells, the classification accuracy of 99.99% with Kappa coefficient of 0.9998 was obtained from Support Vector Machine (SVM) classification algorithm.