Improved U-Net-Like Network for Visual Saliency Detection Based on Pyramid Feature Attention

Figure 2

Detailed structure of the context-aware pyramid feature extraction module. The context-aware feature extraction module takes the high-level features output by the encoder of U-Net-like backbone as input and is composed of three convolutional layers with dilated filters with different dilated rates and one convolutional layer. CFE in this figure stands for context-aware feature extraction module.