CBAM: Convolutional Block Attention Module

TLDR

We propose CBAM, a lightweight attention module for feed‑forward CNNs, and will release its code and models. CBAM sequentially computes channel and spatial attention maps, multiplies them to refine feature maps, and can be inserted into any CNN with minimal overhead, being end‑to‑end trainable and validated on ImageNet‑1K, MS COCO, and VOC 2007. Experiments demonstrate consistent improvements in classification and detection across multiple models, confirming CBAM’s broad applicability.

Abstract

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.