Lung Infection Detection via CT Images and Transfer Learning Techniques in Deep Learning

Abstract

Healthcare systems are battling the global coronavirus epidemic with limited resources, requiring early diagnosis and enhanced tools for pandemic prevention. The computer can aid in diagnosis via computed tomography images like PCR. Deep learning and machine learning are popular methods, and their main contributions are COVID-19 detection and prediction. This work aimed to develop an AI-based early detection strategy for COVID-19 based on computed tomography images. The model was trained and tested using a dataset that includes CT images. The SARS-COV-2 dataset contains 2482 CT images of 210 patients from publicly available sources. The modified model demonstrated encouraging outcomes by greatly enhancing the sensitivity measure (95.82±1.75), which is an essential criterion for accurately detecting instances of COVID-19 infection. In addition, the model generated higher values for the accuracy metric (91.67±1.68), the specificity (88.08±3.72), the precision metric (87.51±3.27), the F1_score (91.43±1.55), and the area under the curve (91.98±1.55). Deep learning techniques significantly facilitate the early detection of COVID-19. Its use has the potential to improve clinical doctors' readiness and the management of pandemics.