Research map / Detection and segmentation / Tracking
Online Metric-Weighted Linear Representations for Robust Visual Tracking
Xi Li, Chunhua Shen, Anthony Dick, Zhongfei Mark Zhang, Yueting Zhuang
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 metric learning to the linear representation (ablation) | tracker with raw pixel features · trace sequence | VOR: 0.07 -> 0.37 | Improved |
| add metric learning to the linear representation (ablation) | tracker with raw pixel features · trace sequence | CLE: 139.30 -> 22.88 | Improved |
| add metric learning (ablation) | tracker with weighted sampling · trace sequence | VOR: 0.51 -> 0.70 | Improved |
| add time-weighted reservoir sampling instead of uniform (ablation) | metric-weighted tracker · trace sequence | VOR: 0.61 -> 0.70 | Improved |
| add time-weighted reservoir sampling (ablation) | metric-weighted tracker · trace sequence | CLE: 15.52 -> 8.65 | Improved |
| add time-weighted reservoir sampling (ablation) | metric-weighted tracker · cubicle sequence | CLE: 4.31 -> 4.31 | No clear effect |
| use step-by-step eigendecomposition instead of none (ablation) | metric learning · trace sequence | VOR: 0.70 -> 0.75 | Improved |
| use final eigendecomposition instead of none (ablation) | metric learning · trace sequence | VOR: 0.70 -> 0.52 | Hurt |
| add structured distance metric learning (Ours+S) | Ours · trellis70 sequence | VOR: 0.78 -> 0.83 | Improved |
| add structured distance metric learning (Ours+S) | Ours · B-Beam sequence | CLE: 4.68 -> 3.52 | Improved |
| use L1 representation instead of metric-weighted linear representation | tracker with raw pixels · race sequence | CLE: 5.18 -> 160.79 | Hurt |
| increase sample buffer size (ablation) | reservoir sampling · five video sequences | CLE and VOR | Improved |
Closest papers by meaning
- Tracking by Animation: Unsupervised Learning of Multi-Object Attentive Trackers (CVPR 2019 · similarity 0.86)
- Tracklet Association by Online Target-Specific Metric Learning and Coherent Dynamics Estimation (TPAMI 2016 · similarity 0.86)
- Release the Power of Online-Training for Robust Visual Tracking (AAAI 2020 · similarity 0.85)
- Video Diffusion Models Excel at Tracking Similar-Looking Objects Without Supervision (NeurIPS 2025 · similarity 0.85)
- VastTrack: Vast Category Visual Object Tracking (NeurIPS 2024 · similarity 0.84)
- Robust Tracking via Mamba-based Context-aware Token Learning (AAAI 2025 · similarity 0.83)
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.
