Channel Attention was first used as a squeeze and excitation block for classification, which generates channel attention maps by using the relationship between the channels.
- Image Super-Resolution Using Very Deep Residual Channel Attention Networks(ECCV).[paper][code]
- CBAM: Convolutional Block Attention Module(ECCV).[paper][keras][code]
- BAM: Bottleneck Attention Module(BMVC).[paper][code]
- (Learning a Discriminative Feature Network for Semantic Segmentation).[paper]
- RCA-U-Net: Residual Channel Attention U-Net for Fast Tissue Quantification in Magnetic Resonance Fingerprinting(MICCAI).[paper]
- Bilinear Attention Networks for Person Retrieval(ICCV).[paper][code]
- DenseNet with Deep Residual Channel-Attention Blocks for Single Image Super Resolution(CVPR Workshop).[paper][code]