Rapid Identification of Pathogens based on MIE Light Scattering and Machine Learning Approach

The detection of pathogens in food is an essential part of food quality control and safety plan. A laser light scattering system designed for rapid and label-free detection of bacterial pathogens. The sample prepared with unknown bacterial contamination and placed inside the system where a laser beam focused on it. When microbial particles pass through the laser beam, light absorbed, refracted and scattered by these particles. The intensity of scattered light measured by an assembly of twelve sensors and features were extracted using power spectrums characteristics from the time domain signal. Different bacterial microbes show different patterns of scattered light depending upon their size, shape, and morphology. The power spectral features used for modeling a classifier for classification using Support Vector Machines (SVM). SVM classifier trained for classification of three bacterial microbes, Enterococcus faecalis, Escherichia coli and Staphylococcus aureus, with resulting average identification accuracies of 98.8%, 79.65%, and 82.3%, respectively. Results indicate that the power spectrum features extraction and SVM classification can achieve significant results in pathogens identification. We believe that the proposed laser-based system has potential for rapid and label-free microbial identification.