Add-Net: Leveraging Additive Attention Mechanism in U-Net Architecture for Precise Liver Tumor Segmentation

Accurate liver tumor segmentation in medical imaging is critical for effective diagnosis and treatment planning in patients with liver diseases. This paper proposes an enhanced U-Net architecture incorporating an additive attention mechanism, referred to as Add-Net, designed to improve segmentation performance on complex medical images. The Add-Net is evaluated on the LiTS dataset, a widely recognized benchmark for liver and tumor segmentation tasks. The integration of the additive attention mechanism allows the network to focus more effectively on relevant features, leading to improved boundary delineation and segmentation accuracy. Comparative analysis demonstrates that the Add-Net model achieves a Dice score of 99.90, outperforming state-of-the-art models. These results suggest that the Add-Net provides a robust solution for liver tumor segmentation in clinical settings, offering significant improvements in segmentation precision and reliability.