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

In this study, we propose a robust methodology for automatic segmentation of infected lung regions in COVID-19 CT scans using convolutional neural networks. The approach is based on a modified U-Net architecture enhanced with attention mechanisms, data augmentation, and postprocessing techniques. It achieved a Dice coefficient of 0.8658 and mean IoU of 0.8316, outperforming other methods. The dataset was sourced from public repositories and augmented for diversity. Results demonstrate superior segmentation performance. Future work includes expanding the dataset, exploring 3D segmentation, and preparing the model for clinical deployment.

💡 Research Summary

The paper addresses the urgent need for reliable, automatic segmentation of COVID‑19 infected lung regions in computed tomography (CT) scans, a task that supports diagnosis, severity assessment, and treatment planning. The authors propose a modified U‑Net architecture that incorporates attention gates, a pre‑trained ResNe​t‑34 encoder, and a suite of advanced loss functions, combined with extensive data augmentation and a post‑processing pipeline.

Data were sourced from publicly available repositories (Coronacases.org, Radiopaedia.org, Zenodo) and consisted of 20 three‑dimensional CT studies, each with dimensions 512 × 512 × 301 voxels and expert‑annotated infection masks. After clipping Hounsfield Units to the clinically relevant range (‑1000 to 1500) and normalising intensities to the