Automatic Detection of Covid-19 with Bidirectional LSTM Network Using Deep Features Extracted from Chest X-ray Images

Abstract

The main aim of this study is to present an effective and distinct deep learning-based model that shows a successful performance in Covid-19 disease which has adversely affects humanity since 2020.Here, the performances of two deep learning architectures on the deep features are compared for Covid-19 disease.<br>In this context, the main contribution of this study is as follows:a) The effectiveness of the Deep Neural Networks (DNN) and Bidirectional Long Short Term Memory (Bi-LSTM) deep learning models were compared comprehensively.b) The performances of the models within the framework of the 5-fold cross-validation technique were verified on the validation datasets.Finally, a highly efficient deep learning model was derived for detecting Covid-19 disease.