mlsanityData to 7 Oct 2026

LF-ViT: Reducing Spatial Redundancy in Vision Transformer for Efficient Image Recognition

arXiv 2402.00033 ↗First posted 8 Jan 2024cs.CV cs.AIcited by 13

Not on the map. Why?

Abstract (arXiv)

The Vision Transformer (ViT) excels in accuracy when handling high-resolution images, yet it confronts the challenge of significant spatial redundancy, leading to increased computational and memory requirements. To address this, we present the Localization and Focus Vision Transformer (LF-ViT). This model operates by strategically curtailing computational demands without impinging on performance. In the Localization phase, a reduced-resolution image is processed; if a definitive prediction remains elusive, our pioneering Neighborhood Global Class Attention (NGCA) mechanism is triggered, effectively identifying and spotlighting class-discriminative regions based on initial findings. Subsequently, in the Focus phase, this designated region is used from the original image to enhance recognition. Uniquely, LF-ViT employs consistent parameters across both phases, ensuring seamless end-to-end optimization. Our empirical tests affirm LF-ViT's prowess: it remarkably decreases Deit-S's FLOPs by 63% and concurrently amplifies throughput twofold. Code of this project is at https://github.com/edgeai1/LF-ViT.git.

Key facts. Cited by 13 papers in the map (data through 7 Oct 2026). Builds on 33 of its 44 references that are in the map. First posted on arXiv 8 Jan 2024.

Builds on

33 of 44 references.

References · 33

  1. 2017Attention Is All You Needcited by 59,932
  2. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
  3. 2021Swin Transformer: Hierarchical Vision Transformer using Shifted Windowscited by 12,082
  4. 2020End-to-End Object Detection with Transformerscited by 7,384
  5. 2021Pyramid Vision Transformer: A Versatile Backbone for Dense Prediction without Convolutionscited by 1,960
  6. 2021Tokens-to-Token ViT: Training Vision Transformers from Scratch on ImageNetcited by 1,096
  7. 2021CvT: Introducing Convolutions to Vision Transformerscited by 946
  8. 2017Revisiting Unreasonable Effectiveness of Data in Deep Learning Eracited by 916
  9. 2021CrossViT: Cross-Attention Multi-Scale Vision Transformer for Image Classificationcited by 806
  10. 2021AdaViT: Adaptive Tokens for Efficient Vision Transformercited by 360
  11. 2017SkipNet: Learning Dynamic Routing in Convolutional Networkscited by 328
  12. 2021LocalViT: Analyzing Locality in Vision Transformerscited by 228
21 more
  1. 2017BlockDrop: Dynamic Inference Paths in Residual Networkscited by 185
  2. 2021Evo-ViT: Slow-Fast Token Evolution for Dynamic Vision Transformercited by 175
  3. 2021Conformer: Local Features Coupling Global Representations for Visual Recognitioncited by 164
  4. 2017Adaptive Neural Networks for Efficient Inferencecited by 161
  5. 2020Rethinking Transformer-based Set Prediction for Object Detectioncited by 156
  6. 2021All Tokens Matter: Token Labeling for Training Better Vision Transformerscited by 144
  7. 2021Not All Images are Worth 16x16 Words: Dynamic Transformers for Efficient Image Recognitioncited by 143
  8. 2021Visformer: The Vision-friendly Transformercited by 137
  9. 2022QuadTree Attention for Vision Transformerscited by 129
  10. 2021IA-RED2: Interpretability-Aware Redundancy Reduction for Vision Transformerscited by 127
  11. 2020Glance and Focus: a Dynamic Approach to Reducing Spatial Redundancy in Image Classificationcited by 99
  12. 2022Wave-ViT: Unifying Wavelet and Transformers for Visual Representation Learningcited by 78
  13. 2019AI Benchmark: All About Deep Learning on Smartphones in 2019cited by 76
  14. 2019Improved Techniques for Training Adaptive Deep Networkscited by 66
  15. 2019LiteEval: A Coarse-to-Fine Framework for Resource Efficient Video Recognitioncited by 65
  16. 2022CF-ViT: A General Coarse-to-Fine Method for Vision Transformercited by 63
  17. 2023SpectFormer: Frequency and Attention is what you need in a Vision Transformercited by 48
  18. 2021Vision Transformer with Progressive Samplingcited by 44
  19. 2023DFormer: Diffusion-guided Transformer for Universal Image Segmentationcited by 19
  20. 2022SaiT: Sparse Vision Transformers through Adaptive Token Pruningcited by 15
  21. 2021Visual Search Asymmetry: Deep Nets and Humans Share Similar Inherent Biasescited by 12

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 1; 2025: 7; 2026: 5.

13 papers.

Sort:

By area:

  1. 2025CATP: Contextually Adaptive Token Pruning for Efficient and Enhanced Multimodal In-Context Learningcited by 18
  2. 2025Accelerating Vision Transformers with Adaptive Patch Sizescited by 12
  3. 2025EnergyFormer: Energy Attention with Fourier Embedding for Hyperspectral Image Classificationcited by 4
  4. 2026DDiT: Dynamic Patch Scheduling for Efficient Diffusion Transformerscited by 3
  5. 2024ESP-PCT: Enhanced VR Semantic Performance through Efficient Compression of Temporal and Spatial Redundancies in Point Cloud Transformerscited by 2
  6. 2025SAC-ViT: Semantic-Aware Clustering Vision Transformer with Early Exitcited by 1
  7. 2025Task-based Loss Functions in Computer Vision: A Comprehensive Reviewcited by 1
  8. 2026Efficiency Follows Global-Local Decouplingcited by 1
  9. 2026Adaptive Volumetric Mechanical Property Fields Invariant to Resolutioncited by 1
  10. 2025A Synthetic-to-Real Dehazing Method based on Domain Unificationcited by 0
  11. 2025Towards an Effective Action-Region Tracking Framework for Fine-grained Video Action Recognitioncited by 0
  12. 2026A Saccade-inspired Approach to Image Classification using Vision Transformer Attention Mapscited by 0
1 more
  1. 2026ProgResViT: Progressive Resolution and Width for Adaptive Vision Transformerscited by 0