Enhanced ResNet-101 Model for Effective Classification of Covid-19 Disease from CXR Images

The recent spread of covid-19 (Corona virus) turned out as a deadliest outbreak in human history, infecting over 636 million people world-wide and claiming an estimated 6.6 million lives. With such a deadly disease spreading around the world at such a fast pace, identification of the covid-19 affected personnel is vital against the spreading of the disease and recovery of the infected individuals. Numerous mechanisms have come forth for covid-19 detection in infected patients. One such method is employing Deep Learning (DL) strategies. Supervised, semi-supervised, and unsupervised models can be built using DL, a subclass of Machine Learning (ML). Artificial neural networks served as inspiration for this. DL can extract features in a hierarchical manner from the data, i.e., low-level feature extraction is done first, followed by mid-level features, and high-level feature extraction is done last. This work expresses the need of DL techniques to apply on Chest Radiograph (CXR) images for effectively identifying the Covid19 infection.