Respiratory Sound Classification for COVID-19 Preliminary Screening

As the COVID-19 virus pandemic continues, we have seen a huge increase in concern for patient diagnosis and treatment or the human respiratory system. Combined with lack of effective treatment thus far, medical resources deem to be insufficient. Thus, an accurate and curated screening process to detect the early onset symptoms of this virus are greatly needed. In this paper, we propose the use of various machine learning techniques to showcase the efficacy of respiratory sound classification, ultimately attempting to improve upon the traditional physician diagnosis procedure. This proposed research work discovers unique classification methodology which offers a competitive alternative to a traditional hands-on approach, and aims to help develop novel data-driven models for identifying COVID-19 patient symptoms in their preliminary screenings during and after this pandemic.