Fig. 4 | Scientific Reports

Fig. 4

From: A 3D medical image segmentation network based on gated attention blocks and dual-scale cross-attention mechanism

Fig. 4

G-SWPA network architecture. The parallel integration of spatial attention and channel attention is achieved through a gating mechanism, which dynamically regulates the impact of both types of attention on feature extraction to enhance the effectiveness of feature information acquisition. Both G_SA and G_CA serve as learnable parameters that can be automatically generated based on current size during calculations, with gradients computed during backpropagation for parameter updates.

Back to article page