Research map / Detection and segmentation / Tracking
HieDG: A Hierarchical Discrete Geometry-Guided Framework for Multi-Animal Tracking
Chenxun Deng, Zhongde Zhang, Ye Yuan, Chengyang Zhang, Yifan Zhang, Bohao Chen, Hongying Yan, Hang Zhou, Hua Han, Xi Chen
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add continuous geometric embeddings (ablation) | MOTIP · BuckTales | HOTA: 45.6 -> 41.3 | Hurt |
| Quantize geometry with two-stage residual codebooks instead of continuous embeddings (ablation) | HieDG* · BuckTales | HOTA: 41.3 -> 52.4 | Improved |
| Quantize geometry with two-stage residual codebooks instead of continuous embeddings (ablation) | HieDG* · BFT | HOTA: 69.5 -> 71.3 | Improved |
| Quantize geometry with two-stage residual codebooks instead of continuous embeddings (ablation) | HieDG* · AnimalTrack | HOTA: 54.6 -> 56.2 | Improved |
| Quantize geometry with two-stage residual codebooks instead of continuous embeddings (ablation) | HieDG* · DanceTrack | HOTA: 68.5 -> 70.3 | Improved |
| Add position cue (ablation) | MOTIP · BFT | HOTA: 69.2 -> 70.6 | Improved |
| Add position, scale and velocity cues together (ablation) | MOTIP · BFT | HOTA: 69.2 -> 71.3 | Improved |
| Add velocity cue only (ablation) | MOTIP · BFT | HOTA: 69.2 -> 69.7 | Improved |
| Add a second residual codebook stage (ablation) | 1-Stage VQ · BFT | HOTA: 70.5 -> 71.3 | Improved |
| Add a third residual codebook stage (ablation) | 2-Stage VQ · BFT | HOTA: 71.3 -> 71.5 | No clear effect |
| Increase codebook dimension from 16 to 64 (ablation) | HieDG · BFT | HOTA: 70.2 -> 71.3 | Improved |
| Increase codebook dimension from 64 to 128 (ablation) | HieDG · BFT | HOTA: 71.3 -> 71.4 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- ByteTrack — ByteTrack: Multi-Object Tracking by Associating Every Detection Box (ECCV 2022) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- NetTrack: Tracking Highly Dynamic Objects with a Net (CVPR 2024 · similarity 0.85)
- BuckTales: A multi-UAV dataset for multi-object tracking and re-identification of wild antelopes (NeurIPS 2024 · similarity 0.85)
- MammAlps: A Multi-view Video Behavior Monitoring Dataset of Wild Mammals in the Swiss Alps (CVPR 2025 · similarity 0.84)
- BEAST3D: animal behavioral analysis and neural encoding from multi-view video via Gaussian splatting (NeurIPS 2026 · similarity 0.84)
- Animal-Bench: Benchmarking Multimodal Video Models for Animal-centric Video Understanding (NeurIPS 2024 · similarity 0.84)
- Pose Splatter: A 3D Gaussian Splatting Model for Quantifying Animal Pose and Appearance (NeurIPS 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
