AI Paper Atlas

Research map / Detection and segmentation / CNN features

SeaFormer++: Squeeze-Enhanced Axial Transformer for Mobile Visual Recognition

Qiang Wan, Zilong Huang, Jiachen Lu, Gang Yu, Li Zhang

IJCV 2025 · Detection and segmentation › CNN features · approach: Plain and high-resolution ViTs · 197 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add detail enhancement kernel to squeeze axial attention (ablation)SeaFormer-T · ADE20K valmIoU: 33.5 -> 35.8Improved
use multiplication instead of addition for enhancement mode (ablation)SeaFormer-T · ADE20K valmIoU: 35.4 -> 35.8Improved
share convolution weights between detail enhancement and squeeze attention (ablation)SeaFormer-T · ADE20K valmIoU: 35.9 -> 35.8No clear effect
use adaptive squeeze and expand instead of mean pooling and broadcast (ablation)SeaFormer-T · ImageNet-1K and ADE20Ktop-1 / mIoU: 67.9 -> 69.8 top-1; 35.8 -> 36.8 mIoUImproved
use max pooling instead of mean pooling for squeeze (ablation)SeaFormer-T · ADE20K valmIoU: 35.8 -> 35.0Hurt
use SEA attention instead of global attention (ablation)SeaFormer-T · ImageNet-1K and ADE20Ktop-1 / mIoU: +1.2% Top1; +1.6 mIoUImproved
replace window attention in Swin with SEA attention (ablation)Swin · ADE20K valmIoU: 45.3 -> 46.5Improved
use Light head instead of Simple head (ablation)SeaFormer-B · ADE20K valmIoU: +0.4 mIoUImproved
use MobileNetV2-based upsample module instead of direct interpolation (ablation)SeaFormer++ KD student · ADE20K valmIoU: 33.7 -> 35.5Improved
use standard conv upsample instead of MobileNetV2 module (ablation)SeaFormer++ KD student · ADE20K valmIoU / latency: 35.5 -> 35.7; 22ms -> 30msMixed
add output similarity loss to classification loss in KD (ablation)SeaFormer++ KD · ADE20K valmIoU: 32.5 -> 33.7Improved
add feature similarity loss (ablation)SeaFormer++ KD · ADE20K valmIoU: 33.7 -> 34.7Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy