A Comparison of COVID-19 Detection using Deep Learning Methods

• Published in: 2023 Second International Conference on Electronics and Renewable Systems (ICEARS) pp. 1345 - 1351
• Main Authors: Kumar, S. Pradeep, Murugan, Suganiya, Rubini, B.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 02-03-2023
• Subjects: Convolutional Neural Network (CNN); Convolutional neural networks; COVID-19; Deep learning; Deep Learning (DL); Neural networks; Predictive models; Renewable energy sources; Training; Tuning; X-ray chest images
• DOI: 10.1109/ICEARS56392.2023.10085367

The recognition of covid-19 is major confront in today's world, specified as sudden increase in spreading of the disease. Hence, identifying this infection in earlier phase facilitates medicinal fields such as doctors, nurses and lab reporters. This article introduces a novel deep learning technique especially Convolutional Neural Network (CNN) by analyzing features in chest input images. Moreover, this proposed Convolutional Neural Network detects the covid-19 disease under several layers and finally performs binary classification that categorizes input images into covid 19 and non-covid patients. Finally, comparisons had made among all models to predict which model diagnose the disease accurately. To evaluate the overall model performance in detection and classification of covid disease, metrics criterias precision, recall and F1-score are evaluated. Validation analysis were completed for quantify the outcomes via performance measures for each model. This proposed comparison attains maximum accuracy of 100% along with least loss as 0.04 that might diminish human inaccuracy in identification procedure.