Throughout the globe a new infection named as coronavirus, that is spreading among human being very fast and intensely. Due to the fast spread of this virus since December 2019, the financial activities across the whole world are deteriorating. There was a lockdown in the whole world because of which the world’s biggest stock markets have collapsed. Unemployment in the whole world has increased in a large number and the trade between the countries stopped. To stop the spread of virus between person to person, the World Health Organization (WHO) has advised the people to adopt the home isolation. The main challenge in this pandemic is to identify the infected people from this virus. The present method which are commonly used are measuring of body temperature and doing blood test. However, body temperature detection and lab testing of the blood is complex and intrusive. The current challenge is to develop some technology to non-intrusively detect the suspected coronavirus patients at crowded places through the COVID alike symptoms of cough, sneezing and flu. Another, challenge to conduct the research on this area is the difficulty to obtain the data set due to limited number of patients to give their consent to be part of the research study. Looking at the efficacy of Artificial Intelligence (AI) in healthcare systems, it is a great challenge for the researchers to develop an AI algorithm which can assist health professionals and government officials to automatically identify and segregate the people having coronavirus symptoms such as cough and flu. Hence, this paper proposes a novel proof of concept system using ML-DCNNet to identify the Coronavirus infected people through facial expression (FE) recognition. The proposed algorithm takes the facial expressions of the people and identifies the facial expressions linked with normal health, cough, sneezing and flu. The data of the facial expressions have been collected through market places, medical clinics and quarantine centers in India. The working of the developed algorithm has been divided into dual stages, at the first stage, the suspected COVID infected patients are classified using Expression-Net on the basis of FEs and in the second stage, intensity level is checked using Intensity-Net to segregate the suspected people with cough, sneezing and flu symptoms. The proposed prototype of ML-DCNN is used to measure the people infected with COVID-19 with their symptoms intensity estimation has been carried out by using the COVID-19 datasets. The proposed system will act as a COVID alert system about the presence of suspected Coronavirus infected people with symptoms of cough, sneezing and flue. It is the first kind of study to analyze the facial expressions and behavioural measures (coughing, sneezing, flu and hand movements). This is study is a proof of concept which can be viable solution in future to detect the suspected COVID patients. However, this needs to be tested on larger dataset. It has been foreseen that the proposed method will demonstrate a distinguished performance as contrast to the situation of the skill methods being used currently.