Pattern recognition of omicron variants from amalgamated multi-focus EEG signals and X-ray images using deep transfer learning

The World Health Organization (WHO) in March 2020 declared an infectious disease caused by the Sars-CoV-2 virus known as COVID-19 as global epidemic. COVID-19 has many variants, the most recent and lethal being the Omicron variant, which has seen an exponential increase in infected cases. The fast spread of Omicron makes diagnosis a key responsibility for health care practitioners. Moreover, recognizing and isolating infected people helps to control the Omicron's spread. For the diagnosis, RT-PCR test is performed which is time consuming and costly. Moreover, in most of the countries the testing is not available for large number of patients due to the unavailability of resources. This research work presents a deep learning-based approach for effectively diagnosis the virus-infected patients using EEG and X-ray images. Effective layered architecture composed of preprocessing, feature extraction (wavelet transformation and efficientNet) and transfer learning based classification has been designed to identify the Omicron patient. From the experimental analysis, it has been concluded that the proposed model produces 96.98 %accuracy with only 12 percent loss and 96 % correct prediction. In order to validate the proposed model, a dataset of EEG Images as well as chest X-rays based images have been collected from online repositories and further classified into 30 % EEG images of normal COVID and 70 % EEG images of Omicron respectively.