Corpus ID: 49864419

BAM: Bottleneck Attention Module

@inproceedings{Park2018BAMBA,
  title={BAM: Bottleneck Attention Module},
  author={Jongchan Park and S. Woo and Joon-Young Lee and In-So Kweon},
  booktitle={BMVC},
  year={2018}
}
Recent advances in deep neural networks have been developed via architecture search for stronger representational power. [...] Key Method Our module infers an attention map along two separate pathways, channel and spatial. We place our module at each bottleneck of models where the downsampling of feature maps occurs. Our module constructs a hierarchical attention at bottlenecks with a number of parameters and it is trainable in an end-to-end manner jointly with any feed-forward models. We validate our BAM…Expand
158 Citations
Auto Learning Attention
  • 2
  • PDF
DCANet: Learning Connected Attentions for Convolutional Neural Networks
  • PDF
CBAM: Convolutional Block Attention Module
  • 1,486
  • PDF
DIANet: Dense-and-Implicit Attention Network
  • 5
  • PDF
Convolutional Neural Network optimization via Channel Reassessment Attention module
  • Highly Influenced
  • PDF
Weight Excitation: Built-in Attention Mechanisms in Convolutional Neural Networks
  • Highly Influenced
  • PDF
Learn to Pay Attention Via Switchable Attention for Image Recognition
  • Highly Influenced
ADNet: Adaptively Dense Convolutional Neural Networks
  • PDF
Channel Attention with Embedding Gaussian Process: A Probabilistic Methodology
  • Highly Influenced
  • PDF
...
1
2
3
4
5
...

References

SHOWING 1-10 OF 50 REFERENCES
CBAM: Convolutional Block Attention Module
  • 1,486
  • PDF
Residual Attention Network for Image Classification
  • 1,365
  • PDF
Deep Pyramidal Residual Networks
  • 321
  • PDF
Squeeze-and-Excitation Networks
  • 3,859
  • PDF
SCA-CNN: Spatial and Channel-Wise Attention in Convolutional Networks for Image Captioning
  • 747
  • PDF
Densely Connected Convolutional Networks
  • 13,041
  • Highly Influential
  • PDF
Aggregated Residual Transformations for Deep Neural Networks
  • 3,611
  • Highly Influential
  • PDF
Identity Mappings in Deep Residual Networks
  • 4,730
  • Highly Influential
  • PDF
Dynamic Filter Networks
  • 471
  • PDF
Going deeper with convolutions
  • 23,575
  • PDF
...
1
2
3
4
5
...