Revisiting the Ordering of Channel and Spatial Attention: A Comprehensive Study on Sequential and Parallel Designs
This work provides scenario-based guidelines for designing attention modules, addressing an empirical gap in fusion strategy selection for researchers and practitioners in computer vision and medical imaging.
The study systematically compared channel-spatial attention fusion strategies across 18 topologies on vision and medical datasets, revealing that optimal designs depend on data scale: few-shot tasks favor 'Channel-Multi-scale Spatial' cascades, medium-scale tasks benefit from parallel learnable fusion, and large-scale tasks perform best with parallel structures using dynamic gating.
Attention mechanisms have become a core component of deep learning models, with Channel Attention and Spatial Attention being the two most representative architectures. Current research on their fusion strategies primarily bifurcates into sequential and parallel paradigms, yet the selection process remains largely empirical, lacking systematic analysis and unified principles. We systematically compare channel-spatial attention combinations under a unified framework, building an evaluation suite of 18 topologies across four classes: sequential, parallel, multi-scale, and residual. Across two vision and nine medical datasets, we uncover a "data scale-method-performance" coupling law: (1) in few-shot tasks, the "Channel-Multi-scale Spatial" cascaded structure achieves optimal performance; (2) in medium-scale tasks, parallel learnable fusion architectures demonstrate superior results; (3) in large-scale tasks, parallel structures with dynamic gating yield the best performance. Additionally, experiments indicate that the "Spatial-Channel" order is more stable and effective for fine-grained classification, while residual connections mitigate vanishing gradient problems across varying data scales. We thus propose scenario-based guidelines for building future attention modules. Code is open-sourced at https://github.com/DWlzm.