CBAM: Convolutional Block Attention Module
By Sanghyun Woo, Jongchan Park, Joon-Young Lee, In-So Kweon. 2018
24,594 citations recorded by the source corpus.
We propose Convolutional Block Attention Module (CBAM), a simple yet effective attention module for feed-forward convolutional neural networks. Given an intermediate feature map, our module sequentially infers attention maps along two separate dimensions, channel and spatial, then the attention maps are multiplied to the input feature map for adaptive feature refinement. Because CBAM is a lightweight and general module, it can be integrated into any CNN architectures seamlessly with negligible overheads and is end-to-end trainable along with base CNNs. We validate our CBAM through extensive experiments on ImageNet-1K, MS COCO detection, and VOC 2007 detection datasets. Our experiments show consistent improvements in classification and detection performances with various models, demonstrating the wide applicability of CBAM. The code and models will be publicly available.
Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Model Architectures
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Dacheng Tao
- Jianfeng Gao
- Quoc V. Le
- Jiashi Feng
- Mohit Bansal
- Trevor Darrell
- Yoshua Bengio
- Alan L. Yuille
In Computer Vision and Pattern Recognition
- Luc Van Gool
- Dacheng Tao
- Ming-Hsuan Yang
- Xiaogang Wang
- Trevor Darrell
- Yi Yang
- Chunhua Shen
- Jitendra Malik
In Model Architectures
Citations in the corpus
- Visual Transformers: Token-based Image Representation and Processing for Computer Vision (2020, methodology): “∆ MACs calculated based on the baseline in [9].”
- Exploring Self-Attention for Image Recognition (2020, methodology): “A number of methods learned to reweight convolutional activations or offset the taps of convolutional kernels [5, 15, 34, 36, 46], thus retaining the basic principle of convolutional feature construction.”
- Local Relation Networks for Image Recognition (2019, background): “In [13], the channel-wise attention are explored and [30] further integrated the channel-wise attention and spatial-wise attention.”
- Attention Augmented Convolutional Networks (2019, methodology): “All attention augmented networks use κ = 2 υ =0 Table 2 benchmarks Attention Augmentation against channel and spatial attention mechanisms BAM [31], CBAM [46] and GALA [28] with channel reduction ratio σ = 16 on the ResNet50 architecture.”
- The Many Faces of Robustness: A Critical Analysis of Out-of-Distribution Generalization (2020, methodology): “When studying self-attention, we employ CBAM [34] and SE [17] modules, two forms of selfattention that help models learn spatially distant dependencies.”
- VOLO: Vision Outlooker for Visual Recognition (2021, background): “CBAM [38] further advances the SE work by introducing spatial attention in CNNs.”
- Towards Real-World Prohibited Item Detection: A Large-Scale X-ray Benchmark (2021, neutral): “We also compare the dependency refinement module with the existing attention mechanisms(e.g. SE and CBAM).”
- Dynamic Neural Networks: A Survey (2021, background): “It is worth noting that the popular soft attention mechanism could also be unified in the framework of dynamic networks, as different channels [20], spatial areas [21] or temporal locations [22] of features are dynamically re-weighted at test time.”
- Squeeze-and-Excitation Networks (2017, background): “Concurrent work provides an interesting study into the combined use of spatial and channel attention around convolutional [58] and bottleneck units [59].”
- Gather-Excite: Exploiting Feature Context in Convolutional Neural Networks (2018, background): “Recent work has developed both powerful and generic parameterised attention modules to allow the system to extract informative signals dynamically [46, 4, 45].”
- BAM: Bottleneck Attention Module (2018, methodology): “A concurrent work, CBAM [43], uses a similar approach using attention modules every convolutional block.”
- Anti-Retroactive Interference for Lifelong Learning (2022, background): “where A ∈ Rs×s denotes the spatial attention obtained by [29], ◦ denotes the Hadamard product, B ∈ Rs×s is the mask for focusing on the background, and”
- Tokens-to-Token ViT: Training Vision Transformers from Scratch on ImageNet (2021, methodology): “Compared with most of the work exploring the global attention on images [1, 42, 13, 39], some works [18, 31] also explore the self-attention in a local patch to reduce the memory and computation cost.”
- VATT: Transformers for Multimodal Self-Supervised Learning from Raw Video, Audio and Text (2021, background): “Early work studied hybrid models consisting of both convolutions and attention modules [94, 100, 40, 111].”
- LambdaNetworks: Modeling Long-Range Interactions Without Attention (2021, background): “Spatial attention Conversely, spatial attention mechanisms, reweigh each position based on signals aggregated from all channels (Xu et al., 2015; Park et al., 2018; Woo et al., 2018).”
- Using Self-Supervised Learning Can Improve Model Robustness and Uncertainty (2019, background): “In addition, since we should like the network to more easily learn shape and compare regions across the whole image, we discovered there is utility in self-attention [Woo et al., 2018] for”
- Adversarial Latent Autoencoders (2020, background): “They allow learning suitable representations of the data for the task at hand, either in a supervised [29, 46, 40, 14, 52], or unsupervised [37, 58, 19, 25, 4, 3] manner.”
- CrossViT: Cross-Attention Multi-Scale Vision Transformer for Image Classification (2021, methodology): “, SENet [18] uses channel-attention, CBAM [41] adds the spatial attention and ECANet [37] proposes an efficient channel attention to further improve SENet.”
- Efficient Attention: Attention with Linear Complexities (2018, background): “On top of SE, CBAM [24] added global max pooling beside global average pooling and an extra spatial attention submodule.”