Recognition and Identification of COVID-19 from Chest X-Rays for Earlier Diagnosis

Early detection is critical for controlling the spread and death of the COVID-19 worldwide pandemic, which affects global health care and lifestyle. The gold standard diagnosis, Reverse Transcription Polymerase Chain Reaction (RT-PCR) is both costly and has a lengthy process time. So more readily available and fast diagnostic methods are necessary. Many researchers proved that COVID-19 data can be effectively detected by Chest X-rays (CXR). However, diagnosing chest X-ray images for COVID-19 is difficult and error-prone. The only way to improve the COVID-19 diagnosis is to use DL algorithms. In this research, a novel deep learning model Convolutional Neural Network and Autoencoder (CNN-AE) tuned by Bat Optimization Algorithm (BOA) is proposed. The proposed approach would identify whether the CXR images were positive or negative for COVID. We received the CXR images from Kaggle. The pre-processing phases in the proposed technique assist with providing relevant information and reducing biased outputs. This is followed by the classification model. We were able to evaluate the performance of our proposed model by using Kaggle datasets. Our proposed model is evaluated using Kaggle datasets and achieved an average accuracy, True Positive Rate (TPR), True Negative Rate (TNR), precision, F1-score, False Positive Rate (FPR), and False Negative Rate (FNR) of 99.28%, 99.42%, 99.14%, 99.42%, 99.28%, 0.85%, and 0.57% respectively. The outcome of the proposed method is compared with the existing and similar research. A comparative study shows that our proposed method outperforms the state-of-art methods.