Audio texture analysis of COVID-19 cough, breath, and speech sounds

•Our proposed model can classify 5 types of cough sounds with an accuracy rate of 71.7%, 5 types of breath sounds with an accuracy rate of 72.2%, and 79.7% of speech sounds. The system offers the highest accuracy rate of 98.9% while performing binary classification on COVID-19 and non-COVID-19 cough sounds.•To our knowledge, this is the first time that the audio textural analysis on COVID-19 cough and breath sounds was done on following five different classes: COVID-19 positive with cough, COVID-19 positive without cough, healthy person with cough, healthy person without cough, and an asthmatic cough.•The proposed work is one of first works in which audio texture is explored to screen COVID-19 sounds i.e. cough, breath, and speech.