AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Vision Transformer With Quadrangle Attention

Qiming Zhang, Jing Zhang, Yufei Xu, Dacheng Tao

TPAMI 2024 · Architectures and efficiency › Attention and transformers · approach: Efficient and robust vision transformers · 111 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 quadrangle attention in place of fixed windows (ablation)ViT-B/16 window attention · ImageNet-1k valtop-1 accuracy: 81.2 -> 82.9Improved
add shear transformation on top of scale and shift (ablation)QFormer_p-B · ImageNet-1k valtop-1 accuracy: 82.3 -> 82.5Improved
add rotation on top of scale, shift, shear (ablation)QFormer_p-B · ImageNet-1k valtop-1 accuracy: 82.5 -> 82.6Improved
add projection on top of scale, shift, shear, rotation (ablation)QFormer_p-B · ImageNet-1k valtop-1 accuracy: 82.6 -> 82.9Improved
use quadrangle attention instead of window attention with 2 full attention layers (ablation)ViT-B/16 · ImageNet-1k valtop-1 accuracy: 82.5 -> 83.0Improved
use quadrangle attention instead of window attention with 6 full attention layers (ablation)ViT-B/16 · ImageNet-1k valtop-1 accuracy: 82.9 -> 83.3Improved
compose basic transformations instead of predicting the matrix directly (ablation)QA in ViT-B/16 · ImageNet-1k valtop-1 accuracy: 82.7 -> 82.9Improved
quadrangle attention instead of shifted window attentionViT-B/16 · ImageNet-1k valtop-1 accuracy: 82.0 -> 82.9Improved
add the regularization term (lambda 0 to 0.1) (ablation)QFormer_p-B · ImageNet-1k valtop-1 accuracy: diverged -> 82.9Improved
increase regularization weight lambda from 1 to 10 (ablation)QFormer_p-B · ImageNet-1k valtop-1 accuracy: 82.9 -> 82.7Hurt
replace window attention with quadrangle attentionViT-B + UPerNet · ADE20k valmIoU: 39.7 -> 43.6Improved
replace Swin-T windows with quadrangle attentionSwin-T Mask RCNN 1x · COCO valbox mAP: 43.7 -> 45.9Improved

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