Research map / Detection and segmentation / Tracking
A Novel Performance Evaluation Methodology for Single-Target Trackers
Matej Kristan, Jiří Matas, Aleš Leonardis, Tomas Vojir, Roman P. Pflugfelder, Gustavo Fernandez, Georg Nebehay, Fatih Porikli, Luka Čehovin Zajc
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 statistical and practical difference tests to ranking (ablation) | VOT ranking · VOT2014 robustness, attribute-normalized | rank variance: 0.34 -> 0.09 | Improved |
| add difference tests (ablation) | VOT ranking · VOT2014 robustness, sequence-pooled | rank variance: 0.1 -> 0.07 | Improved |
| add difference tests (ablation) | VOT ranking · VOT2014 accuracy, sequence-pooled | rank variance: 0.11 -> 0.1 | Improved |
| add difference tests (ablation) | VOT ranking · VOT2014 accuracy, attribute-normalized | rank variance: 0.31 -> 0.26 | Improved |
| exclude burn-in frames (10) from accuracy (ablation) | tracker accuracy · VOT2014 | average accuracy: 0.5534 -> 0.5455 | Improved |
| increase N_skip from 1 to 9 (ablation) | tracker robustness · VOT2014 | failures (DSST): 1.32 -> 1.00 | No clear effect |
| add re-initialization after failure (ablation) | tracker evaluation · VOT2014 K=5 | average overlap (CT): 0.23 -> 0.43 | Improved |
| use per-frame visual attribute annotation (vs per-sequence) | attribute evaluation · VOT2014 | bias of per-attribute estimate: 0.67 -> 0.09 | Hurt |
| use per-sequence/per-frame annotation M_GLA instead of M_PFA | estimator · VOT2014 | variance: (1+beta)/(1+eta) lower variance | Mixed |
Closest papers by meaning
- Beyond Standard Benchmarks: Parameterizing Performance Evaluation in Visual Object Tracking (ICCV 2017 · similarity 0.88)
- SOTVerse: A User-Defined Task Space of Single Object Tracking (IJCV 2023 · similarity 0.87)
- LaSOT: A High-quality Large-scale Single Object Tracking Benchmark (IJCV 2020 · similarity 0.86)
- HOTA: A Higher Order Metric for Evaluating Multi-object Tracking (IJCV 2020 · similarity 0.86)
- Is This Tracker On? A Benchmark Protocol for Dynamic Tracking (NeurIPS 2025 · similarity 0.84)
- DriveTrack: A Benchmark for Long-Range Point Tracking in Real-World Videos (CVPR 2024 · 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.
