Research map / Graph learning / Knowledge and temporal graphs
Learning Patch-Based Dynamic Graph for Visual Tracking
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 learned patch weights to the tracker (ablation) | Struck-based tracker · OTB100 | PR: 0.795 -> 0.865 | Helped |
| add patch weights (ablation) | Struck-based tracker · OTB100 | SR: 0.558 -> 0.586 | Helped |
| use learned low-rank sparse dynamic graph instead of 8-neighbor graph (ablation) | Struck-based tracker · OTB100 | PR: 0.829 -> 0.865 | Helped |
| use dynamic graph instead of 8-neighbor graph (ablation) | Struck-based tracker · OTB100 | SR: 0.569 -> 0.586 | Helped |
| add overfitting-avoidance term (xi/2 ||w||^2) in graph objective (ablation) | Struck-based tracker · OTB100 | PR: 0.840 -> 0.865 | Helped |
| add overfitting-avoidance term (ablation) | Struck-based tracker · OTB100 | SR: 0.574 -> 0.586 | Helped |
| add patch-weighted graph representation to Struck | Struck · OTB100 OPE | PR: 0.640 -> 0.865 | Helped |
| set beta to 2 and xi to 2 | Ours · OTB100 | PR/SR: beta=2: 0.7% lower PR, 0.1% better SR; xi=2: 0.1% lower both | No clear effect |
| use patch-weighted color/gradient features on low-resolution sequences | Ours vs HCF · OTB100 LR attribute | PR: 0.787 (HCF) vs 0.732 (Ours) | Hurt |
Closest papers by meaning
- Multi-Scene Generalized Trajectory Global Graph Solver with Composite Nodes for Multiple Object Tracking (AAAI 2024 · similarity 0.88)
- Video Diffusion Models Excel at Tracking Similar-Looking Objects Without Supervision (NeurIPS 2025 · similarity 0.87)
- DiffusionTrack: Diffusion Model for Multi-Object Tracking (AAAI 2024 · similarity 0.87)
- Track-On: Transformer-based Online Point Tracking with Memory (ICLR 2025 · similarity 0.86)
- DeTrack: In-model Latent Denoising Learning for Visual Object Tracking (NeurIPS 2024 · similarity 0.86)
- OneTracker: Unifying Visual Object Tracking with Foundation Models and Efficient Tuning (CVPR 2024 · similarity 0.86)
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.