Attention-Enhanced U-Net for Accurate Segmentation of COVID-19 Infected Lung Regions in CT Scans

Automated Segmentation of COVID - 19 Infected Lung Regions 2 Abstract In this study, the focus is on developing a robust methodology for automatic segmentation of infected lung regions in COVID - 19 CT scans utilizing advanced CNNs. The proposed model is based on a modified U - Net architecture w ith attention mechanisms, data augmentation, and postprocessing techniques, achieving high segmentation accuracy and boundary precision. The dataset was sourced from publicly available repositories, processed, and augmented to increase its diversity and ge neralizability. The approach was evaluated quantitatively, resulting in a Dice coefficient of 0.8658 and mean IoU of 0.8316. The proposed model is compared to existing methods through comparative analysis, clearly demonstrating its superiority in handling data variability and achieving precise segmentation.