The COVID-19 virus outbreak has exceeded our expectations and shattered all previous records for virus outbreaks. The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) causes progressive respiratory failure and severe alveolar damage might be deadly. During the pandemic, to curb the virus' spread and ease the strain on the healthcare systems, these has arisen an imperative for swift and precise detection of COVID-19 through computer-aided diagnosis. This paper aims to study the effects of the original and five image enhancement techniques which are Modified Global Contrast stretching (MGCS), Adaptive Gamma Correction with Weighting Distribution (AGCWD), Lowlight (LL), Multi-scale Retinex 2 (MSR2), and Contrast Enhancement using Heat Conduction Matrix (CEHCM) on chest X-ray (CXR) images on the classification process. As a matter of fact, to attain accurate and quick COVID-19 detection, a standard convolutional neural network (CNN) and long short-term memory (LSTM) were developed. A total of 15000 CXR images consisting of COVID-19, normal, and pneumonia were collected from various data repositories to implement this study. The experimental result shows the best classification performance of the CNN-LSTM model is achieved when the system is fed with CXR images enhanced by the lowlight (LL) image enhancement technique, which achieves accuracy, sensitivity, specificity, precision, and F1-score of 99.65%, 99.80%, 99.95%, 99.90%, and 99.85%, respectively.
