Helium-327/DCLA-UNet-3D: 首次发布
Abstract
DCLA-UNet introduces a Dynamic Cross Layer Attention (DCLA) mechanism that dynamically selects the most relevant features from different layers of the network. This attention mechanism enhances the feature representation capability of U-Net for medical image segmentation tasks, particularly for brain tumor segmentation in BraTS datasets.
Publication details
- Journal / source
- Zenodo (CERN European Organization for Nuclear Research)
- Publication date
- 7 January 2026