mlsanityData to 7 Oct 2026

When Fish Look Alike: Tracking Identities with Dual-branch Elasticity

arXiv 2607.26412 ↗First posted 29 Jul 2026cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

Tracking dense, homogeneous targets like schooling fish remains a major challenge for multiple object tracking due to extreme inter-individual homogeneity, severe physical clustering, and rapid non-rigid deformations. While heavy-backbone separated detection and embedding trackers like SU-T push accuracy boundaries using complex Re-Identification networks, their computational overhead prohibits edge deployment. Furthermore, these modules often fail when appearance features degrade under severe occlusions. To overcome this, we propose Tracking Identities with Dual-branch Elasticity (TIDE). Bypassing expensive appearance cues, TIDE utilizes the Adaptive Geometric Correspondence IoU, an association mechanism leveraging spatial and structural consistency to robustly handle complex morphological variations. Crucially, TIDE introduces system-level deployment elasticity, decoupling the algorithmic pipeline from strict hardware constraints. Evaluations on the MFT-Edge benchmark demonstrate that our Lightweight L-branch achieves a competitive HOTA of 28.43 using merely 20.47G FLOPs. This represents a 38.7-fold computational reduction compared to upper bounds like SU-T, directly facilitating real-time edge deployment. Concurrently, our Scalable S-branch establishes a 29.98 HOTA, successfully bridging the gap between high-precision cloud analysis and efficient edge tracking. The dataset and codes are released at https://vranlee.github.io/TIDE/.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 24 of its 32 references that are in the map. First posted on arXiv 29 Jul 2026.

Builds on

24 of 32 references.

References · 24

  1. 2016Feature Pyramid Networks for Object Detectioncited by 5,021
  2. 2019Generalized Intersection over Union: A Metric and A Loss for Bounding Box Regressioncited by 1,361
  3. 2017Multi-Task Learning Using Uncertainty to Weigh Losses for Scene Geometry and Semanticscited by 1,345
  4. 2016Simple Online and Realtime Trackingcited by 862
  5. 2021ByteTrack: Multi-Object Tracking by Associating Every Detection Boxcited by 812
  6. 2020HOTA: A Higher Order Metric for Evaluating Multi-Object Trackingcited by 546
  7. 2021TrackFormer: Multi-Object Tracking with Transformerscited by 515
  8. 2020FairMOT: On the Fairness of Detection and Re-Identification in Multiple Object Trackingcited by 489
  9. 2020Tracking Objects as Pointscited by 483
  10. 2019Distance-IoU Loss: Faster and Better Learning for Bounding Box Regressioncited by 454
  11. 2017Deep Layer Aggregationcited by 436
  12. 2022Observation-Centric SORT: Rethinking SORT for Robust Multi-Object Trackingcited by 345
12 more
  1. 2019Towards Real-Time Multi-Object Trackingcited by 311
  2. 2020Enhancing Geometric Factors in Model Learning and Inference for Object Detection and Instance Segmentationcited by 98
  3. 2023Hybrid-SORT: Weak Cues Matter for Online Multi-Object Trackingcited by 66
  4. 2022QDTrack: Quasi-Dense Similarity Learning for Appearance-Only Multiple Object Trackingcited by 61
  5. 2023OVTrack: Open-Vocabulary Multiple Object Trackingcited by 53
  6. 2024Efficient Deformable ConvNets: Rethinking Dynamic and Sparse Operator for Vision Applicationscited by 49
  7. 2022Simple Cues Lead to a Strong Multi-Object Trackercited by 44
  8. 2021TransCenter: Transformers with Dense Representations for Multiple-Object Trackingcited by 42
  9. 2024Magic Tokens: Select Diverse Tokens for Multi-modal Object Re-Identificationcited by 34
  10. 2024Salience DETR: Enhancing Detection Transformer with Hierarchical Salience Filtering Refinementcited by 30
  11. 2023Object-Centric Multiple Object Trackingcited by 15
  12. 2025When Trackers Date Fish: A Benchmark and Framework for Underwater Multiple Fish Trackingcited by 2

Built on by

Not cited yet.