Cough parameter is an important measure indicator for COVID-19 disease, spread by SARS COV-2 Virus and an indicator for 100 different diseases. Through coughing the virus can spread from person to person in very eager manner. For example, when an infected individual coughs, respiratory droplets containing the virus can be released into the air, which can affect the person who comes into contact these droplets. The collected samples of cough sound helps in better screening of COVID-19 patients and their health severity. In order to have transparency regarding normal cough sound and infected cough sound(COVID-19),the cough audio data samples are collected from patients around the world, so it become possible to differentiate between COVID-19 positive, negative results and severe cough. The proposed AI model using VGG-19 mainly recognise the patterns in cough categorize the patient type is COVID infected or disinfected. Determining the top 13 highlights using power-hungry sequential direct decision (SFS) calculations and a VGG-19 classifier arrives at 0.94. With the continuous measures taken to stop the wide spread of SARC-COV 2 virus infection wherever today, and against comparative sicknesses in our hypothetical arrangement, with its minimal expense and ease of use, can play a very crucial for pre-diagnosis and analysis. We believe that hacking sounds recorded on mobile phones, or a web interface can be used to measure COVID-19 would be easy way to monitor the patients remotely.

Artificial learning, Neural Networks, SARC-COV 2 virus. VGG-19 model Classifier.