Swin Transformer: Hierarchical Vision Transformer using Shifted Windows

Swin Transformer is a hierarchical vision transformer using shifted windows for efficient local self-attention and linear image complexity, achieving state-of-the-art results on ImageNet, COCO, and ADE20K.

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

Published Oct 1, 202133,075 citationsPaper ↗

91%

OverallMust read

OverallMust readVote to see the scoreThe exact score shows once you've voted, so every vote is your own call. The first half of each home page shelf shows its scores.

Readers

–

Only vote on papers you've read. Sign in with GitHub to vote.

AI panel18/21reviewers recommend it

lenient 5/5

medium 10/11

strict 3/5

AI panel?Vote to see what the 20 AI reviewers said

Panel consensus

Swin Transformer delivers decisive vision backbone gains through hierarchical shifted windows and linear image complexity, though its windowed local attention and training comparisons leave open whether it is a true Transformer breakthrough or an elaborate CNN…

Abstract

This paper presents a new vision Transformer, called Swin Transformer, that capably serves as a general-purpose backbone for computer vision. Challenges in adapting Transformer from language to vision arise from differences between the two domains, such as large variations in the scale of visual entities and the high resolution of pixels in images compared to words in text. To address these differences, we propose a hierarchical Transformer whose representation is computed with Shifted windows. The shifted windowing scheme brings greater efficiency by limiting self-attention computation to non-overlapping local windows while also allowing for cross-window connection. This hierarchical architecture has the flexibility to model at various scales and has linear computational complexity with respect to image size. These qualities of Swin Transformer make it compatible with a broad range of vision tasks, including image classification (87.3 top-1 accuracy on ImageNet-1K) and dense prediction tasks such as object detection (58.7 box AP and 51.1 mask AP on COCO test-dev) and semantic segmentation (53.5 mIoU on ADE20K val). Its performance surpasses the previous state-of-the-art by a large margin of +2.7 box AP and +2.6 mask AP on COCO, and +3.2 mIoU on ADE20K, demonstrating the potential of Transformer-based models as vision backbones. The hierarchical design and the shifted window approach also prove beneficial for all-MLP architectures. The code and models are publicly available at https://github.com/microsoft/Swin-Transformer.