An Unsupervised Prediction Model for Salmonella Detection with Hyperspectral Microscopy: A Multi-Year Validation

Hyperspectral microscope images (HMIs) have been previously explored as a tool for the early and rapid detection of common foodborne pathogenic bacteria. A robust unsupervised classification approach to differentiate bacterial species with potential single cell sensitivity is needed real-world application, in order confirm identity bacteria isolated from food product. Here, one-class soft independent modelling class analogy (SIMCA) was used determine if individual cells are Salmonella positive or negative. The model constructed validated spectral library built over five years, containing 13 serotypes 14 non-Salmonella pathogens. An image processing method designed take less than one minute paired prediction algorithm resulted an overall accuracy 95.4%, 0.97, specificity 0.92. SIMCA’s only achieved after incorporating multiple established. These results demonstrate HMI sensitive presumptive screening method, moving towards (<8 h) (<1 identification matrices.