mlsanityData to 7 Oct 2026

Global Filter Networks for Image Classification

arXiv 2107.00645 ↗First posted 1 Jul 2021cs.CV cs.AI cs.LGcited by 257

Not on the map. Why?

Abstract (arXiv)

Recent advances in self-attention and pure multi-layer perceptrons (MLP) models for vision have shown great potential in achieving promising performance with fewer inductive biases. These models are generally based on learning interaction among spatial locations from raw data. The complexity of self-attention and MLP grows quadratically as the image size increases, which makes these models hard to scale up when high-resolution features are required. In this paper, we present the Global Filter Network (GFNet), a conceptually simple yet computationally efficient architecture, that learns long-term spatial dependencies in the frequency domain with log-linear complexity. Our architecture replaces the self-attention layer in vision transformers with three key operations: a 2D discrete Fourier transform, an element-wise multiplication between frequency-domain features and learnable global filters, and a 2D inverse Fourier transform. We exhibit favorable accuracy/complexity trade-offs of our models on both ImageNet and downstream tasks. Our results demonstrate that GFNet can be a very competitive alternative to transformer-style models and CNNs in efficiency, generalization ability and robustness. Code is available at https://github.com/raoyongming/GFNet

Key facts. Cited by 257 papers in the map (data through 7 Oct 2026). Builds on 35 of its 61 references that are in the map. First posted on arXiv 1 Jul 2021.

Builds on

35 of 61 references.

References · 35

  1. 2017Attention Is All You Needcited by 59,932
  2. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  3. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
  4. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
  5. 2021Swin Transformer: Hierarchical Vision Transformer using Shifted Windowscited by 12,082
  6. 2020End-to-End Object Detection with Transformerscited by 7,384
  7. 2014Explaining and Harnessing Adversarial Examplescited by 7,287
  8. 2017Mask R-CNNcited by 7,259
  9. 2019EfficientNet: Rethinking Model Scaling for Convolutional Neural Networkscited by 6,415
  10. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,326
  11. 2020Training data-efficient image transformers & distillation through attentioncited by 4,560
  12. 2019Benchmarking Neural Network Robustness to Common Corruptions and Perturbationscited by 2,857
23 more
  1. 2020Fourier Neural Operator for Parametric Partial Differential Equationscited by 2,806
  2. 2016Aggregated Residual Transformations for Deep Neural Networkscited by 2,540
  3. 2021Pyramid Vision Transformer: A Versatile Backbone for Dense Prediction without Convolutionscited by 1,960
  4. 2019Do ImageNet Classifiers Generalize to ImageNet?cited by 1,507
  5. 2021MLP-Mixer: An all-MLP Architecture for Visioncited by 1,441
  6. 2019Natural Adversarial Examplescited by 1,401
  7. 2020Rethinking Semantic Segmentation from a Sequence-to-Sequence Perspective with Transformerscited by 1,368
  8. 2017Random Erasing Data Augmentationcited by 1,235
  9. 2021Tokens-to-Token ViT: Training Vision Transformers from Scratch on ImageNetcited by 1,096
  10. 2016Deep Networks with Stochastic Depthcited by 973
  11. 2020Designing Network Design Spacescited by 954
  12. 2021CvT: Introducing Convolutions to Vision Transformerscited by 946
  13. 2021DynamicViT: Efficient Vision Transformers with Dynamic Token Sparsificationcited by 849
  14. 2021Going deeper with Image Transformerscited by 759
  15. 2019Panoptic Feature Pyramid Networkscited by 644
  16. 2020FDA: Fourier Domain Adaptation for Semantic Segmentationcited by 540
  17. 2021Transformer Trackingcited by 412
  18. 2021Pay Attention to MLPscited by 367
  19. 2021FNet: Mixing Tokens with Fourier Transformscited by 343
  20. 2021ResMLP: Feedforward networks for image classification with data-efficient trainingcited by 338
  21. 2020Are we done with ImageNet?cited by 329
  22. 2017CirCNN: Accelerating and Compressing Deep Neural Networks Using Block-CirculantWeight Matricescited by 53
  23. 2018FALCON: A Fourier Transform Based Approach for Fast and Secure Convolutional Neural Network Predictionscited by 20

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 15; 2022: 45; 2023: 55; 2024: 53; 2025: 42; 2026: 47.

257 papers.

Sort:

By area:

  1. 2022A ConvNet for the 2020scited by 3,545
  2. 2022FEDformer: Frequency Enhanced Decomposed Transformer for Long-term Series Forecastingcited by 1,200
  3. 2024Vision Mamba: Efficient Visual Representation Learning with Bidirectional State Space Modelcited by 1,051
  4. 2021Resolution-robust Large Mask Inpainting with Fourier Convolutionscited by 730
  5. 2023One Fits All:Power General Time Series Analysis by Pretrained LMcited by 615
  6. 2021DenseCLIP: Language-Guided Dense Prediction with Context-Aware Promptingcited by 530
  7. 2021MetaFormer Is Actually What You Need for Visioncited by 523
  8. 2022MAXIM: Multi-Axis MLP for Image Processingcited by 301
  9. 2021Adaptive Fourier Neural Operators: Efficient Token Mixers for Transformerscited by 237
  10. 2022Towards Multi-spatiotemporal-scale Generalized PDE Modelingcited by 190
  11. 2021Adaptive Token Sampling For Efficient Vision Transformerscited by 181
  12. 2022Scaling Up Your Kernels to 31x31: Revisiting Large Kernel Design in CNNscited by 155
38 more
  1. 2022FiLM: Frequency improved Legendre Memory Model for Long-term Time Series Forecastingcited by 147
  2. 2022MetaFormer Baselines for Visioncited by 131
  3. 2021CycleMLP: A MLP-like Architecture for Dense Predictioncited by 126
  4. 2022HorNet: Efficient High-Order Spatial Interactions with Recursive Gated Convolutionscited by 103
  5. 2022Filter-enhanced MLP is All You Need for Sequential Recommendationcited by 96
  6. 2022Clifford Neural Layers for PDE Modelingcited by 91
  7. 2023PoseFormerV2: Exploring Frequency Domain for Efficient and Robust 3D Human Pose Estimationcited by 91
  8. 2022Conv2Former: A Simple Transformer-Style ConvNet for Visual Recognitioncited by 81
  9. 2024SiMBA: Simplified Mamba-Based Architecture for Vision and Multivariate Time seriescited by 81
  10. 2021Intriguing Findings of Frequency Selection for Image Deblurringcited by 80
  11. 2023Joint Token Pruning and Squeezing Towards More Aggressive Compression of Vision Transformerscited by 79
  12. 2022Fast Vision Transformers with HiLo Attentioncited by 78
  13. 2024Wavelet Convolutions for Large Receptive Fieldscited by 74
  14. 2022LargeKernel3D: Scaling up Kernels in 3D Sparse CNNscited by 71
  15. 2024TSLANet: Rethinking Transformers for Time Series Representation Learningcited by 71
  16. 2021Hire-MLP: Vision MLP via Hierarchical Rearrangementcited by 66
  17. 2023UniRepLKNet: A Universal Perception Large-Kernel ConvNet for Audio, Video, Point Cloud, Time-Series and Image Recognitioncited by 64
  18. 2021An Image Patch is a Wave: Phase-Aware Vision MLPcited by 61
  19. 2023Frequency-Aware Transformer for Learned Image Compressioncited by 56
  20. 2023SpectFormer: Frequency and Attention is what you need in a Vision Transformercited by 48
  21. 2024Frequency-aware Feature Fusion for Dense Image Predictioncited by 48
  22. 2023Adaptive Frequency Filters As Efficient Global Token Mixerscited by 45
  23. 2024UltraLight VM-UNet: Parallel Vision Mamba Significantly Reduces Parameters for Skin Lesion Segmentationcited by 45
  24. 2024A Survey on Transformer Compressioncited by 44
  25. 2023Toeplitz Neural Network for Sequence Modelingcited by 43
  26. 2022Deep Frequency Filtering for Domain Generalizationcited by 41
  27. 2024Frequency-Adaptive Dilated Convolution for Semantic Segmentationcited by 34
  28. 2024H-vmunet: High-order Vision Mamba UNet for Medical Image Segmentationcited by 34
  29. 2021A Battle of Network Structures: An Empirical Study of CNN, Transformer, and MLPcited by 32
  30. 2023Out-of-distributional risk bounds for neural operators with applications to the Helmholtz equationcited by 32
  31. 2022Unraveling Attention via Convex Duality: Analysis and Interpretations of Vision Transformerscited by 30
  32. 2024Visual Fourier Prompt Tuningcited by 30
  33. 2022DynaMixer: A Vision MLP Architecture with Dynamic Mixingcited by 28
  34. 2023ALOFT: A Lightweight MLP-like Architecture with Dynamic Low-frequency Transform for Domain Generalizationcited by 28
  35. 2022Attribute Surrogates Learning and Spectral Tokens Pooling in Transformers for Few-shot Learningcited by 27
  36. 2023Mobile Foundation Model as Firmwarecited by 25
  37. 2021S2-MLPv2: Improved Spatial-Shift MLP Architecture for Visioncited by 24
  38. 2022Sequencer: Deep LSTM for Image Classificationcited by 23

50 most cited of 257 All 257 (JSON)