AI Paper Atlas

Research map / Detection and segmentation / CNN features

Swin Transformer: Hierarchical Vision Transformer Using Shifted Windows

Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, Baining Guo

ICCV 2021 · Detection and segmentation › CNN features · approach: Plain and high-resolution ViTs · 38,712 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 shifted window partitioning between consecutive blocks (ablation)Swin-T · ImageNet-1Ktop-1 accuracy: 80.2 -> 81.3Improved
add shifted window partitioning between consecutive blocks (ablation)Swin-T · COCO (Cascade Mask R-CNN)box AP: 47.7 -> 50.5Improved
add shifted window partitioning between consecutive blocks (ablation)Swin-T · ADE20KmIoU: 43.3 -> 46.1Improved
add relative position bias instead of no position encoding (ablation)Swin-T · ImageNet-1Ktop-1 accuracy: 80.1 -> 81.3Improved
add relative position bias instead of no position encoding (ablation)Swin-T · ADE20KmIoU: 43.8 -> 46.1Improved
use relative position bias instead of absolute position embedding (ablation)Swin-T · COCO (Cascade Mask R-CNN)box AP: 49.0 -> 50.5Improved
add absolute position embedding on top of relative bias (ablation)Swin-T · ADE20KmIoU: 46.1 -> 44.0Hurt
add absolute position embedding without relative bias (ablation)Swin-T · ImageNet-1Ktop-1 accuracy: 80.1 -> 80.5Improved
replace sliding window self-attention with shifted windows (ablation)Swin-T · ImageNet-1Ktop-1 accuracy: 81.4 -> 81.3No clear effect
use cyclic-shift batch computation instead of padding (ablation)Swin-T · V100 throughputFPS: 670 -> 755Improved
use shifted windows instead of sliding windows (kernel) (ablation)Swin-T · V100 throughputFPS: 488 -> 755Improved
replace Performer attention with shifted-window attention (ablation)Swin-T · ImageNet-1Ktop-1 accuracy: 79.0 -> 81.3Improved

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