Research Article Efficient Framework for Detection of COVID-19 Omicron and Delta Variants Based on Two Intelligent Phases of CNN Models Mustafa Ghaderzadeh , 1 Mohammad Amir Eshraghi , 2 Farkhondeh Asadi , 3 Azamossadat Hosseini , 3 Ramezan Jafari , 4 Davood Bashash , 5 and Hassan Abolghasemi 6 1 Student Research Committee, Department and Faculty of Health Information Technology and Management School of Allied Medical Sciences, Shahid Beheshti University of Medical Sciences, Tehran, Iran 2 School of Electrical Engineering, Iran University of Science and Technology, Tehran, Iran 3 Department of Health Information Technology and Management, School of Allied Medical Sciences, Shahid Beheshti University of Medical Sciences, Tehran, Iran 4 Department of Radiology, Baqiyatallah University of Medical Sciences, Tehran, Iran 5 Department of Hematology and Blood Banking, School of Allied Medical Sciences, Shahid Beheshti University of Medical Sciences, Tehran, Iran 6 Pediatric Congenital Hematologic Disorders Research Center, Shahid Beheshti University of Medical Sciences, Tehran, Iran Correspondence should be addressed to Farkhondeh Asadi; asadifar@sbmu.ac.ir and Azamossadat Hosseini; souhosseini@sbmu.ac.ir Received 23 December 2021; Revised 10 March 2022; Accepted 16 March 2022; Published 21 April 2022 Academic Editor: Cristiana Corsi Copyright © 2022 Mustafa Ghaderzadeh et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Introduction. While the COVID-19 pandemic was waning in most parts of the world, a new wave of COVID-19 Omicron and Delta variants in Central Asia and the Middle East caused a devastating crisis and collapse of health-care systems. As the diagnostic methods for this COVID-19 variant became more complex, health-care centers faced a dramatic increase in patients. Thus, the need for less expensive and faster diagnostic methods led researchers and specialists to work on improving diagnostic testing. Method. Inspired by the COVID-19 diagnosis methods, the latest and most efficient deep learning algorithms in the field of extracting X-ray and CT scan image features were used to identify COVID-19 in the early stages of the disease. Results. We presented a general framework consisting of two models which are developed by convolutional neural network (CNN) using the concept of transfer learning and parameter optimization. The proposed phase of the framework was evaluated on the test dataset and yielded remarkable results and achieved a detection sensitivity, specificity, and accuracy of 0.99, 0.986, and 0.988, for the first phase and 0.997, 0.9976, and 0.997 for the second phase, respectively. In all cases, the whole framework was able to successfully classify COVID-19 and non-COVID-19 cases from CT scans and X-ray images. Conclusion. Since the proposed framework was based on two deep learning models that used two radiology modalities, it was able to significantly assist radiologists in detecting COVID-19 in the early stages. The use of models with this feature can be considered as a powerful and reliable tool, compared to the previous models used in the past pandemics. 1. Introduction Two years after the emergence of coronavirus, while many countries were experiencing an overall decline in cases, COVID-19 variants have led to a spike in deaths and hospi- talizations in some other countries. Since late November 2021, a new wave of uncertainty and panic of the pandemic has spread around the world, as the number of cases of coro- navirus 2019 (COVID-19) has increased dramatically, espe- cially in the southern part of Africa, Europe, and East of Asia. A new type of SARS-CoV-2, B.1.1.529, was identified by the World Health Organization as a type of concern Hindawi Computational and Mathematical Methods in Medicine Volume 2022, Article ID 4838009, 10 pages https://doi.org/10.1155/2022/4838009