Abstract:
Early efforts to predict the spread of disease have been traced back to the 14th century. Most notable in historical disease prediction is the Black Death which engulfed Europe from 1347 to 1351, killing an estimated 200 million people. Attempts to predict the spread of disease have, until recently, relied essentially on weather predictions coupled with the emergence of known vectors of a given pathogen. Recent software developments have allowed computer models to make use of predictive analysis of disease trends to help isolate the most likely path for disease expansion. Data from hospitals and clinics, combined with key word search of social media and air travel data add additional variables to the prediction process.
The advent of the present COVID-19 Pandemic clearly demonstrates the urgent need for airborne pathogen detection methods, and associated decision analysis regarding the virulence and the pathogenicity of an infectious agent. The motivation of the dissertation was to create a crowd-sourced means of reference pathogens image templates, necessary for deep learning algorithm development for computer vision identification of unknown pathogens. Utilizing deep learning analytical techniques as a tool to augment traditional chemical signature information to yield near real time identification of airborne pathogens, thereby yielding improved decision matrix accuracy outcomes.
