Cutting-Edge Multi-Task Model: Unveiling Covid-19 Through Fusion of Image Processing Algorithms

COVID-19 poses a global health challenge with over 458 million confirmed cases and 6.1 million deaths worldwide. Accurate diagnosis of lung infections through CT scans is crucial for effective treatment. However, manual diagnosis is time-consuming and subjective. This study proposed a multi-task model that combines classification and segmentation tasks using an encoder-decoder architecture based on the U-net model and image processing algorithms to improve the image quality. The model utilizes two datasets: the "MedSeg" and the "COVID-19 CT Lung and Infection Segmentation." For both datasets, the combination of median and opening operations achieves high accuracy of 0.97 and 0.96, excellent image quality with PSNR values of 36.94 and 36.87, and strong structural similarity with SSIM scores of 0.95 and 0.94. The proposed model outperforms traditional approaches, demonstrating superior segmentation and classification results. These findings contribute to the advancement of medical image analysis techniques, potentially enhancing diagnostic accuracy in healthcare settings.