The rapid global proliferation of COVID-19 poses significant challenges to the timely effective identification and monitoring of infected cases. Traditional diagnostic methods, such as Polymerase Chain Reaction, are often supplemented or even replaced by radiological techniques, due to their limitations in detecting COVID-19. Therefore, medical imaging modalities, notably chest X-ray images and computed tomography scans, play an indispensable role in the diagnostic detection process for COVID-19. Because chest X-ray imaging has a lower informational sensitivity compared to computed tomography scans, chest X-ray imaging complicates data classifications and results in diminished accuracy for deep learning models. Our study leverages advanced image enhancement techniques to delineate regions of interest, by achieving remarkable classification accuracies of 100% for the VGG-16 model from the chest X-ray images and 94% from the computed tomography scans, as well as 94% for AlexNet model from the chest X-rays images and 80% from the computed tomography scans. These classification accuracies underscore the considerable promise of image enhancement techniques in computer-aided diagnosis; thereby, they are offering significant implications for healthcare applications.