mlsanityData to 14 Oct 2026

DARTH: Holistic Test-time Adaptation for Multiple Object Tracking

arXiv 2310.01926 ↗First posted 3 Oct 2023cs.CV cs.AIcited by 15

Not on the map. Why?

Abstract (arXiv)

Multiple object tracking (MOT) is a fundamental component of perception systems for autonomous driving, and its robustness to unseen conditions is a requirement to avoid life-critical failures. Despite the urge of safety in driving systems, no solution to the MOT adaptation problem to domain shift in test-time conditions has ever been proposed. However, the nature of a MOT system is manifold - requiring object detection and instance association - and adapting all its components is non-trivial. In this paper, we analyze the effect of domain shift on appearance-based trackers, and introduce DARTH, a holistic test-time adaptation framework for MOT. We propose a detection consistency formulation to adapt object detection in a self-supervised fashion, while adapting the instance appearance representations via our novel patch contrastive loss. We evaluate our method on a variety of domain shifts - including sim-to-real, outdoor-to-indoor, indoor-to-outdoor - and substantially improve the source model performance on all metrics. Code: https://github.com/mattiasegu/darth.

Key facts. Cited by 15 papers in the map (data through 14 Oct 2026). Builds on 58 of its 80 references that are in the map. First posted on arXiv 3 Oct 2023.

Builds on

58 of 80 references.

References · 58

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2015Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networkscited by 9,639
  3. 2018Representation Learning with Contrastive Predictive Codingcited by 7,777
  4. 2017Mask R-CNNcited by 7,277
  5. 2016Feature Pyramid Networks for Object Detectioncited by 5,036
  6. 2015You Only Look Once: Unified, Real-Time Object Detectioncited by 4,852
  7. 2020Bootstrap your own latent: A new approach to self-supervised Learningcited by 4,750
  8. 2016The Cityscapes Dataset for Semantic Urban Scene Understandingcited by 4,444
  9. 2018Unsupervised Feature Learning via Non-Parametric Instance-level Discriminationcited by 1,845
  10. 2018Unsupervised Representation Learning by Predicting Image Rotationscited by 1,591
  11. 2018Group Normalizationcited by 1,524
  12. 2016Unsupervised Learning of Visual Representations by Solving Jigsaw Puzzlescited by 1,382
38 more
  1. 2016Colorful Image Colorizationcited by 1,114
  2. 2017Simple Online and Realtime Tracking with a Deep Association Metriccited by 910
  3. 2016Simple Online and Realtime Trackingcited by 868
  4. 2021ByteTrack: Multi-Object Tracking by Associating Every Detection Boxcited by 823
  5. 2020HOTA: A Higher Order Metric for Evaluating Multi-Object Trackingcited by 552
  6. 2016MOT16: A Benchmark for Multi-Object Trackingcited by 532
  7. 2022Continual Test-Time Domain Adaptationcited by 525
  8. 2021TrackFormer: Multi-Object Tracking with Transformerscited by 518
  9. 2018Domain Adaptive Faster R-CNN for Object Detection in the Wildcited by 443
  10. 2020Improving robustness against common corruptions by covariate shift adaptationcited by 431
  11. 2017Semantic Foggy Scene Understanding with Synthetic Datacited by 403
  12. 2019Tracking without bells and whistlescited by 382
  13. 2021MOTR: End-to-End Multiple-Object Tracking with Transformercited by 381
  14. 2020TransTrack: Multiple Object Tracking with Transformercited by 371
  15. 2018A Survey of Unsupervised Deep Domain Adaptationcited by 361
  16. 2022Observation-Centric SORT: Rethinking SORT for Robust Multi-Object Trackingcited by 346
  17. 2014Discriminative Unsupervised Feature Learning with Exemplar Convolutional Neural Networkscited by 300
  18. 2018CrowdHuman: A Benchmark for Detecting Human in a Crowdcited by 284
  19. 2018Strong-Weak Distribution Alignment for Adaptive Object Detectioncited by 270
  20. 2022Contrastive Test-Time Adaptationcited by 261
  21. 2019Libra R-CNN: Towards Balanced Learning for Object Detectioncited by 255
  22. 2016Learning Representations for Automatic Colorizationcited by 251
  23. 2020Evaluating Prediction-Time Batch Normalization for Robustness under Covariate Shiftcited by 236
  24. 2020Quasi-Dense Similarity Learning for Multiple Object Trackingcited by 220
  25. 2018Cross-Domain Weakly-Supervised Object Detection through Progressive Domain Adaptationcited by 218
  26. 2021DanceTrack: Multi-Object Tracking in Uniform Appearance and Diverse Motioncited by 217
  27. 2020Unbiased Mean Teacher for Cross-domain Object Detectioncited by 159
  28. 2022SHIFT: A Synthetic Driving Dataset for Continuous Multi-Task Domain Adaptationcited by 146
  29. 2019Exploring Object Relation in Mean Teacher for Cross-Domain Detectioncited by 137
  30. 2017Detect to Track and Track to Detectcited by 131
  31. 2021The Norm Must Go On: Dynamic Unsupervised Domain Adaptation by Normalizationcited by 131
  32. 2020Harmonizing Transferability and Discriminability for Adapting Object Detectorscited by 126
  33. 2020Every Pixel Matters: Center-aware Feature Alignment for Domain Adaptive Object Detectorcited by 107
  34. 2021Unsupervised Domain Adaptation of Object Detectors: A Surveycited by 93
  35. 2019Self-Training and Adversarial Background Regularization for Unsupervised Domain Adaptive One-Stage Object Detectioncited by 83
  36. 2016Learning by tracking: Siamese CNN for robust target associationcited by 79
  37. 2020A Free Lunch for Unsupervised Domain Adaptive Object Detection without Source Datacited by 79
  38. 2019Automatic adaptation of object detectors to new domains using self-trainingcited by 66

50 most cited of 58 All 58 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2023: 1; 2024: 8; 2025: 4; 2026: 2.

15 papers.

Sort:

By area:

  1. 2024Matching Anything by Segmenting Anythingcited by 50
  2. 2024Towards Generalizable Multi-Object Trackingcited by 23
  3. 2024Samba: Synchronized Set-of-Sequences Modeling for Multiple Object Trackingcited by 21
  4. 2024SLAck: Semantic, Location, and Appearance Aware Open-Vocabulary Trackingcited by 7
  5. 2025Semantic Library Adaptation: LoRA Retrieval and Fusion for Open-Vocabulary Semantic Segmentationcited by 7
  6. 2024Is Multiple Object Tracking a Matter of Specialization?cited by 5
  7. 2023COOLer: Class-Incremental Learning for Appearance-Based Multiple Object Trackingcited by 4
  8. 2024Walker: Self-supervised Multiple Object Tracking by Walking on Temporal Appearance Graphscited by 4
  9. 2024Is a Pure Transformer Effective for Separated and Online Multi-Object Tracking?cited by 2
  10. 2025ViLAaD: Enhancing "Attracting and Dispersing'' Source-Free Domain Adaptation with Vision-and-Language Modelcited by 2
  11. 2026Dual-level Adaptation for Multi-Object Tracking: Building Test-Time Calibration from Experience and Intuitioncited by 2
  12. 2025Towards Reliable Test-Time Adaptation: Style Invariance as a Correctness Likelihoodcited by 1
3 more
  1. 2024CrowdTransfer: Enabling Crowd Knowledge Transfer in AIoT Communitycited by 0
  2. 2025EvoPSF: Online Evolution of Autonomous Driving Models via Planning-State Feedbackcited by 0
  3. 2026Cross-Domain Tracker Adaptation Without Target-Domain Labels via Vision-Language Agentscited by 0