seUNet-Attention Based COVID-19 Lung Infection Segmentation Model

This paper addresses the challenge of accurately segmenting COVID-19 lung infections in CT images, crucial for disease diagnosis and treatment. It proposes an improved seUNet-Attention model, enhancing the efficiency and accuracy of segmentation. The model, equipped with an attention mechanism, excels in identifying lung infection features, especially in low-contrast images. Experiments on public datasets show the model's superiority over existing methods, with improved detail preservation and edge segmentation of lesion areas. This advancement significantly supports clinical decision-making in diagnosing and treating COVID-19 lung infections.