mlsanityData to 7 Oct 2026

AiATrack: Attention in Attention for Transformer Visual Tracking

arXiv 2207.09603 ↗First posted 20 Jul 2022cs.CVcited by 137

Not on the map. Why?

Abstract (arXiv)

Transformer trackers have achieved impressive advancements recently, where the attention mechanism plays an important role. However, the independent correlation computation in the attention mechanism could result in noisy and ambiguous attention weights, which inhibits further performance improvement. To address this issue, we propose an attention in attention (AiA) module, which enhances appropriate correlations and suppresses erroneous ones by seeking consensus among all correlation vectors. Our AiA module can be readily applied to both self-attention blocks and cross-attention blocks to facilitate feature aggregation and information propagation for visual tracking. Moreover, we propose a streamlined Transformer tracking framework, dubbed AiATrack, by introducing efficient feature reuse and target-background embeddings to make full use of temporal references. Experiments show that our tracker achieves state-of-the-art performance on six tracking benchmarks while running at a real-time speed.

Key facts. Cited by 137 papers in the map (data through 7 Oct 2026). Builds on 54 of its 68 references that are in the map. First posted on arXiv 20 Jul 2022.

Builds on

54 of 68 references.

References · 54

  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. 2014Microsoft COCO: Common Objects in Contextcited by 17,498
  5. 2020End-to-End Object Detection with Transformerscited by 7,384
  6. 2016Layer Normalizationcited by 5,579
  7. 2020Deformable DETR: Deformable Transformers for End-to-End Object Detectioncited by 3,410
  8. 2017Non-local Neural Networkscited by 2,290
  9. 2019Generalized Intersection over Union: A Metric and A Loss for Bounding Box Regressioncited by 1,361
  10. 2021ByteTrack: Multi-Object Tracking by Associating Every Detection Boxcited by 812
  11. 2018CCNet: Criss-Cross Attention for Semantic Segmentationcited by 717
  12. 2016Fully-Convolutional Siamese Networks for Object Trackingcited by 566
38 more
  1. 2021TrackFormer: Multi-Object Tracking with Transformerscited by 515
  2. 2018GOT-10k: A Large High-Diversity Benchmark for Generic Object Tracking in the Wildcited by 471
  3. 2018SiamRPN++: Evolution of Siamese Visual Tracking with Very Deep Networkscited by 445
  4. 2018LaSOT: A High-quality Benchmark for Large-scale Single Object Trackingcited by 424
  5. 2021Transformer Trackingcited by 412
  6. 2020TransTrack: Multiple Object Tracking with Transformercited by 370
  7. 2021Learning Spatio-Temporal Transformer for Visual Trackingcited by 348
  8. 2019Learning Discriminative Model Prediction for Trackingcited by 344
  9. 2018TrackingNet: A Large-Scale Dataset and Benchmark for Object Tracking in the Wildcited by 342
  10. 2019Attention on Attention for Image Captioningcited by 289
  11. 2018ATOM: Accurate Tracking by Overlap Maximizationcited by 285
  12. 2020Probabilistic Regression for Visual Trackingcited by 231
  13. 2021Associating Objects with Transformers for Video Object Segmentationcited by 225
  14. 2021Transformer Meets Tracker: Exploiting Temporal Context for Robust Visual Trackingcited by 220
  15. 2020Ocean: Object-aware Anchor-free Trackingcited by 212
  16. 2018Acquisition of Localization Confidence for Accurate Object Detectioncited by 174
  17. 2018Neighbourhood Consensus Networkscited by 173
  18. 2020How Much Position Information Do Convolutional Neural Networks Encode?cited by 171
  19. 2019Deeper and Wider Siamese Networks for Real-Time Visual Trackingcited by 156
  20. 2017Need for Speed: A Benchmark for Higher Frame Rate Object Trackingcited by 137
  21. 2021Learning Target Candidate Association to Keep Track of What Not to Trackcited by 131
  22. 2020LaSOT: A High-quality Large-scale Single Object Tracking Benchmarkcited by 130
  23. 2020Know Your Surroundings: Exploiting Scene Information for Object Trackingcited by 119
  24. 2021SSTVOS: Sparse Spatiotemporal Transformers for Video Object Segmentationcited by 106
  25. 2020Deformable Siamese Attention Networks for Visual Object Trackingcited by 97
  26. 2020Efficient Neighbourhood Consensus Networks via Submanifold Sparse Convolutionscited by 92
  27. 2021STMTrack: Template-free Visual Tracking with Space-time Memory Networkscited by 91
  28. 2020Alpha-Refine: Boosting Tracking Performance by Precise Bounding Box Estimationcited by 81
  29. 2020High-Performance Long-Term Tracking with Meta-Updatercited by 79
  30. 2022TransVOD: End-to-End Video Object Detection with Spatial-Temporal Transformerscited by 78
  31. 2021CATs: Cost Aggregation Transformers for Visual Correspondencecited by 77
  32. 2019Learning the Model Update for Siamese Trackerscited by 75
  33. 2021HiFT: Hierarchical Feature Transformer for Aerial Trackingcited by 72
  34. 2020Graph Attention Trackingcited by 70
  35. 2021Joint Inductive and Transductive Learning for Video Object Segmentationcited by 61
  36. 2019Deep Learning for Visual Tracking: A Comprehensive Surveycited by 60
  37. 2021Convolutional Hough Matching Networkscited by 43
  38. 2020Correspondence Networks with Adaptive Neighbourhood Consensuscited by 42

50 most cited of 54 All 54 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 4; 2023: 38; 2024: 42; 2025: 34; 2026: 19.

137 papers.

Sort:

By area:

  1. 2023Emergent Correspondence from Image Diffusioncited by 437
  2. 2024ODTrack: Online Dense Temporal Token Learning for Visual Trackingcited by 129
  3. 2023Generalized Relation Modeling for Transformer Trackingcited by 93
  4. 2024SAMURAI: Adapting Segment Anything Model for Zero-Shot Visual Tracking with Motion-Aware Memorycited by 91
  5. 2023Robust Object Modeling for Visual Trackingcited by 80
  6. 2024Autoregressive Queries for Adaptive Tracking with Spatio-TemporalTransformerscited by 76
  7. 2023Event Stream-based Visual Object Tracking: A High-Resolution Benchmark Dataset and A Novel Baselinecited by 73
  8. 2023HIPTrack: Visual Tracking with Historical Promptscited by 71
  9. 2024Tracking Meets LoRA: Faster Training, Larger Model, Stronger Performancecited by 70
  10. 2023Single-Model and Any-Modality for Video Object Trackingcited by 64
  11. 2024OneTracker: Unifying Visual Object Tracking with Foundation Models and Efficient Tuningcited by 64
  12. 2023Towards Unified Token Learning for Vision-Language Trackingcited by 62
38 more
  1. 2024State Space Model for New-Generation Network Alternative to Transformers: A Surveycited by 59
  2. 2023Exploring Lightweight Hierarchical Vision Transformers for Efficient Visual Trackingcited by 53
  3. 2022Revisiting Color-Event based Tracking: A Unified Network, Dataset, and Metriccited by 51
  4. 2024Explicit Visual Prompts for Visual Object Trackingcited by 50
  5. 2023Frame-Event Alignment and Fusion Network for High Frame Rate Trackingcited by 43
  6. 2023Cross-modal Orthogonal High-rank Augmentation for RGB-Event Transformer-trackerscited by 43
  7. 2024SDSTrack: Self-Distillation Symmetric Adapter Learning for Multi-Modal Visual Object Trackingcited by 43
  8. 2023CiteTracker: Correlating Image and Text for Visual Trackingcited by 36
  9. 2024Exploring Enhanced Contextual Information for Video-Level Object Trackingcited by 36
  10. 2025Towards Universal Modal Tracking with Online Dense Temporal Token Learningcited by 30
  11. 2023Towards Real-World Visual Tracking with Temporal Contextscited by 26
  12. 2023Swift Parameter-free Attention Network for Efficient Super-Resolutioncited by 24
  13. 2025SPMTrack: Spatio-Temporal Parameter-Efficient Fine-Tuning with Mixture of Experts for Scalable Visual Trackingcited by 23
  14. 2025Exploring Decoupled Spatio-Temporal Consistency Learning and Self-Prompting Evolution for Self-Supervised Trackingcited by 23
  15. 2023DropMAE: Learning Representations via Masked Autoencoders with Spatial-Attention Dropout for Temporal Matching Taskscited by 22
  16. 2024OmniVid: A Generative Framework for Universal Video Understandingcited by 22
  17. 2024Mamba-FETrack: Frame-Event Tracking via State Space Modelcited by 21
  18. 2023Unified Sequence-to-Sequence Learning for Single- and Multi-Modal Visual Object Trackingcited by 20
  19. 2023OmniTracker: Unifying Object Tracking by Tracking-with-Detectioncited by 19
  20. 2025MangaNinja: Line Art Colorization with Precise Reference Followingcited by 17
  21. 2023Detect Any Shadow: Segment Anything for Video Shadow Detectioncited by 16
  22. 2023LiteTrack: Layer Pruning with Asynchronous Feature Extraction for Lightweight and Efficient Visual Trackingcited by 16
  23. 2024WebUOT-1M: Advancing Deep Underwater Object Tracking with A Million-Scale Benchmarkcited by 15
  24. 2024Robust Tracking via Mamba-based Context-aware Token Learningcited by 15
  25. 2025What You Have is What You Track: Adaptive and Robust Multimodal Trackingcited by 15
  26. 2024Cross-modulated Attention Transformer for RGBT Trackingcited by 14
  27. 2025Do It Yourself: Learning Semantic Correspondence from Pseudo-Labelscited by 13
  28. 2024Reading Relevant Feature from Global Representation Memory for Visual Object Trackingcited by 12
  29. 2022Beyond SOT: Tracking Multiple Generic Objects at Oncecited by 11
  30. 2025Adaptive Perception for Unified Visual Multi-modal Object Trackingcited by 11
  31. 2025UncTrack: Reliable Visual Object Tracking with Uncertainty-Aware Prototype Memory Networkcited by 11
  32. 2023Transformers in Single Object Tracking: An Experimental Surveycited by 10
  33. 2023360VOT: A New Benchmark Dataset for Omnidirectional Visual Object Trackingcited by 10
  34. 2024eMoE-Tracker: Environmental MoE-based Transformer for Robust Event-guided Object Trackingcited by 10
  35. 2025ATCTrack: Aligning Target-Context Cues with Dynamic Target States for Robust Vision-Language Trackingcited by 10
  36. 2023Beyond Visual Cues: Synchronously Exploring Target-Centric Semantics for Vision-Language Trackingcited by 9
  37. 2024TENet: Targetness Entanglement Incorporating with Multi-Scale Pooling and Mutually-Guided Fusion for RGB-E Object Trackingcited by 9
  38. 2024Improving Visual Object Tracking through Visual Promptingcited by 9

50 most cited of 137 All 137 (JSON)