COVID-19 classification using CNN-BiLSTM based on chest X-ray images

• Published in: Bulletin of Electrical Engineering and Informatics
• Main Author: Cahyani D.E.; Hariadi A.D.; Setyawan F.F.; Gumila L.; Setumin S.
• Format: Article
• Language: English
• Published: Institute of Advanced Engineering and Science 2023
• Online Access: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85146441039&doi=10.11591%2feei.v12i3.4848&partnerID=40&md5=3749dc63a16d5b399e9cabb71ae6b2fc

Cases of the COVID-19 virus continue to spread still needs to be considered even though we have entered the post-pandemic era. Rapid identification of COVID-19 cases is necessary to prevent the virus from spreading further. This study developed a chest X-ray-based (CXR) COVID-19 classification for COVID-19 detection using the convolutional neural network-bidirectional long short-term memory (CNN-BiLSTM) combination model and compared the CNN-BiLSTM combination model with CNN models. The CNN models used in this study are the transfer learning models, namely Resnet50, VGG19, InceptionV3, Xception, and AlexNet. This research classifies CXR into three groups: COVID-19, normal, and viral pneumonia. In comparison to other models, the Resnet50-BiLSTM model is the most accurate and hence the best. The accuracy of the Resnet50-BiLSTM model was 98.48%. The model that obtains the next highest accuracy i.e Resnet50, VGG19-BiLSTM, VGG19, InceptionV3-BiLSTM, InceptionV3, Xception-BiLSTM, Xception, AlexNet-BiLSTM, and AlexNet. In this study, precision, recall, and F1-measure are also employed to demonstrate that Resnet50-BiLSTM achieves the highest value compared to other approaches. When compared to previous studies, this study enhances classification performance results. © 2023, Institute of Advanced Engineering and Science. All rights reserved.