Instant Diagnosis Using Raman Spectroscopy and Generative Adversarial Networks: A Blood-Based Study on Seasonal Flu, COVID-19, and Dengue

Abstract

Rapid detection of infectious diseases like COVID-19, flu, and dengue is crucial for healthcare professionals preparing for contagious outbreaks. Given the constant mutations in viruses and the recurring emergence of threats like Nipah and Zika, there is an urgent demand for a technology capable of distinguishing between infections that share similar symptoms. In this paper, we utilize laser-based Raman scattered signals from a drop of dried blood plasma, combined with generative artificial intelligence, to provide a rapid and precise diagnosis. Our optimized model exhibits exceptional performance, yielding high predictive scores of 96%, 98%, and 100% for flu, COVID-19, and dengue, respectively. The proposed Raman spectroscopic analysis, with a rapid turnaround time, can ensure a near-accurate diagnosis and proper quarantining of highly infectious cases. Furthermore, the potential extension of our method to include other viral diseases offers an alternative to the challenge of developing different diagnostic kits for each disease.