mlsanityData to 14 Oct 2026

Tracking Objects as Pixel-wise Distributions

arXiv 2207.05518 ↗First posted 12 Jul 2022cs.CV cs.AI cs.LGcited by 37

Not on the map. Why?

Abstract (arXiv)

Multi-object tracking (MOT) requires detecting and associating objects through frames. Unlike tracking via detected bounding boxes or tracking objects as points, we propose tracking objects as pixel-wise distributions. We instantiate this idea on a transformer-based architecture, P3AFormer, with pixel-wise propagation, prediction, and association. P3AFormer propagates pixel-wise features guided by flow information to pass messages between frames. Furthermore, P3AFormer adopts a meta-architecture to produce multi-scale object feature maps. During inference, a pixel-wise association procedure is proposed to recover object connections through frames based on the pixel-wise prediction. P3AFormer yields 81.2% in terms of MOTA on the MOT17 benchmark -- the first among all transformer networks to reach 80% MOTA in literature. P3AFormer also outperforms state-of-the-arts on the MOT20 and KITTI benchmarks.

Key facts. Cited by 37 papers in the map (data through 14 Oct 2026). Builds on 67 of its 82 references that are in the map. First posted on arXiv 12 Jul 2022.

Builds on

67 of 82 references.

References · 67

  1. 2017Attention Is All You Needcited by 60,361
  2. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  3. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  4. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
  5. 2014Microsoft COCO: Common Objects in Contextcited by 17,608
  6. 2021Swin Transformer: Hierarchical Vision Transformer using Shifted Windowscited by 12,139
  7. 2015Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networkscited by 9,639
  8. 2020End-to-End Object Detection with Transformerscited by 7,414
  9. 2017mixup: Beyond Empirical Risk Minimizationcited by 5,183
  10. 2020Deformable DETR: Deformable Transformers for End-to-End Object Detectioncited by 3,422
  11. 2016DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFscited by 3,241
  12. 2021SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformerscited by 2,940
38 more
  1. 2021Masked-attention Mask Transformer for Universal Image Segmentationcited by 2,351
  2. 2021Pyramid Vision Transformer: A Versatile Backbone for Dense Prediction without Convolutionscited by 1,962
  3. 2021Vision Transformers for Dense Predictioncited by 1,803
  4. 2020YOLOv4: Optimal Speed and Accuracy of Object Detectioncited by 1,630
  5. 2021Per-Pixel Classification is Not All You Need for Semantic Segmentationcited by 1,152
  6. 2021YOLOX: Exceeding YOLO Series in 2021cited by 1,071
  7. 2015FlowNet: Learning Optical Flow with Convolutional Networkscited by 1,059
  8. 2017Simple Online and Realtime Tracking with a Deep Association Metriccited by 910
  9. 2016Simple Online and Realtime Trackingcited by 868
  10. 2021ByteTrack: Multi-Object Tracking by Associating Every Detection Boxcited by 823
  11. 2020HOTA: A Higher Order Metric for Evaluating Multi-Object Trackingcited by 552
  12. 2021TrackFormer: Multi-Object Tracking with Transformerscited by 518
  13. 2020FairMOT: On the Fairness of Detection and Re-Identification in Multiple Object Trackingcited by 490
  14. 2020Tracking Objects as Pointscited by 483
  15. 2022MaxViT: Multi-Axis Vision Transformercited by 476
  16. 2018PVNet: Pixel-wise Voting Network for 6DoF Pose Estimationcited by 385
  17. 2019Tracking without bells and whistlescited by 382
  18. 2021MOTR: End-to-End Multiple-Object Tracking with Transformercited by 381
  19. 2020TransTrack: Multiple Object Tracking with Transformercited by 371
  20. 2020MOT20: A benchmark for multi object tracking in crowded scenescited by 357
  21. 2019DenseFusion: 6D Object Pose Estimation by Iterative Dense Fusioncited by 355
  22. 2022V2X-ViT: Vehicle-to-Everything Cooperative Perception with Vision Transformercited by 320
  23. 2018CrowdHuman: A Benchmark for Detecting Human in a Crowdcited by 284
  24. 2020Quasi-Dense Similarity Learning for Multiple Object Trackingcited by 220
  25. 2017Flow-Guided Feature Aggregation for Video Object Detectioncited by 177
  26. 2017CityPersons: A Diverse Dataset for Pedestrian Detectioncited by 172
  27. 2014Multiple Object Tracking: A Literature Reviewcited by 166
  28. 2021MHFormer: Multi-Hypothesis Transformer for 3D Human Pose Estimationcited by 151
  29. 2021HRFormer: High-Resolution Transformer for Dense Predictioncited by 149
  30. 2020MOTChallenge: A Benchmark for Single-Camera Multiple Target Trackingcited by 136
  31. 2017Detect to Track and Track to Detectcited by 131
  32. 2020Memory Enhanced Global-Local Aggregation for Video Object Detectioncited by 128
  33. 2021Learning to Track with Object Permanencecited by 119
  34. 2020Rethinking the competition between detection and ReID in Multi-Object Trackingcited by 101
  35. 2021Exploiting Temporal Contexts with Strided Transformer for 3D Human Pose Estimationcited by 94
  36. 2021TransMOT: Spatial-Temporal Graph Transformer for Multiple Object Trackingcited by 93
  37. 2022TransVOD: End-to-End Video Object Detection with Spatial-Temporal Transformerscited by 78
  38. 2021AutoMix: Unveiling the Power of Mixup for Stronger Classifierscited by 69

50 most cited of 67 All 67 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2022: 6; 2023: 14; 2024: 6; 2025: 6; 2026: 5.

37 papers.

Sort:

By area:

  1. 2022Tip-Adapter: Training-free Adaption of CLIP for Few-shot Classificationcited by 362
  2. 2023V2V4Real: A Real-world Large-scale Dataset for Vehicle-to-Vehicle Cooperative Perceptioncited by 155
  3. 2022MOTRv2: Bootstrapping End-to-End Multi-Object Tracking by Pretrained Object Detectorscited by 148
  4. 2022AvatarPoser: Articulated Full-Body Pose Tracking from Sparse Motion Sensingcited by 110
  5. 2023MotionTrack: Learning Robust Short-term and Long-term Motions for Multi-Object Trackingcited by 64
  6. 2022Bridging the Domain Gap for Multi-Agent Perceptioncited by 60
  7. 2023SparseTrack: Multi-Object Tracking by Performing Scene Decomposition based on Pseudo-Depthcited by 48
  8. 2023DiffusionTrack: Diffusion Model For Multi-Object Trackingcited by 40
  9. 2022Where is my Wallet? Modeling Object Proposal Sets for Egocentric Visual Query Localizationcited by 24
  10. 2023Video-kMaX: A Simple Unified Approach for Online and Near-Online Video Panoptic Segmentationcited by 24
  11. 2024NetTrack: Tracking Highly Dynamic Objects with a Netcited by 22
  12. 2023ByteTrackV2: 2D and 3D Multi-Object Tracking by Associating Every Detection Boxcited by 20
25 more
  1. 2023Bridging the Gap Between End-to-end and Non-End-to-end Multi-Object Trackingcited by 20
  2. 2023DeNoising-MOT: Towards Multiple Object Tracking with Severe Occlusionscited by 12
  3. 2023Collaborative Tracking Learning for Frame-Rate-Insensitive Multi-Object Trackingcited by 10
  4. 2023Beyond Visual Cues: Synchronously Exploring Target-Centric Semantics for Vision-Language Trackingcited by 9
  5. 2024MCTR: Multi Camera Tracking Transformercited by 8
  6. 2024TrackSSM: A General Motion Predictor by State-Space Modelcited by 6
  7. 2022End-to-end View Synthesis via NeRF Attentioncited by 5
  8. 2025TQD-Track: Temporal Query Denoising for 3D Multi-Object Trackingcited by 4
  9. 2025PapMOT: Exploring Adversarial Patch Attack against Multiple Object Trackingcited by 4
  10. 2023NLOS Dies Twice: Challenges and Solutions of V2X for Cooperative Perceptioncited by 3
  11. 2023OccluTrack: Rethinking Awareness of Occlusion for Enhancing Multiple Pedestrian Trackingcited by 3
  12. 2024Multi-Granularity Language-Guided Training for Multi-Object Trackingcited by 2
  13. 2025When Trackers Date Fish: A Benchmark and Framework for Underwater Multiple Fish Trackingcited by 2
  14. 2023SVT: Supertoken Video Transformer for Efficient Video Understandingcited by 1
  15. 2024IP-MOT: Instance Prompt Learning for Cross-Domain Multi-Object Trackingcited by 1
  16. 2025To New Beginnings: A Survey of Unified Perception in Autonomous Vehicle Softwarecited by 1
  17. 2025Motion-Aware Transformer for Multi-Object Trackingcited by 1
  18. 2023S3Track: Self-supervised Tracking with Soft Assignment Flowcited by 0
  19. 2024An Approximate Dynamic Programming Framework for Occlusion-Robust Multi-Object Trackingcited by 0
  20. 2025DM3T: Harmonizing Modalities via Diffusion for Multi-Object Trackingcited by 0
  21. 2026VSD-MOT: End-to-End Multi-Object Tracking in Low-Quality Video Scenes Guided by Visual Semantic Distillationcited by 0
  22. 2026Multi-view Crowd Tracking Transformer with View-Ground Interactions Under Large Real-World Scenescited by 0
  23. 2026CalibFree: Self-Supervised View Feature Separation for Calibration-Free Multi-Camera Multi-Object Trackingcited by 0
  24. 2026WorldBagel: Uncovering the Power of Unified Multimodal Models for Vision-Language-Action-World Modelingcited by 0
  25. 2026Harnessing LLMs as Agents: What Does It Cost?cited by 0