Research map / Detection and segmentation / Tracking
Siamese network tracking
Track objects by matching templates with Siamese networks, attention, and region proposals.
34 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2025) · an approach in Tracking, keyword cluster “visual tracking · visual object · siamese”
State History Benchmarks Tried Baselines Preprints Who Export
Where it stands
Few recent reports The tracked benchmarks have at most one report from 2025–26 that could beat the old record, too few to say whether they are still moving.
4 papers in 2023–24 and 1 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 6.0% to 1.2%.
Open questions the data points to
- Papers report SiamRPN on VOT-2016 (EAO) anywhere from 0.344 to 0.464: comparisons against it depend on whose number you take.
- Papers report TransT on GOT-10k (SR0.75) anywhere from 60.9 to 68.2: comparisons against it depend on whose number you take.
No longer reported by papers here: VOT2018 (EAO, last in 2021); VOT2016 (EAO, last in 2020); VOT2015 (EAO, last in 2019); OTB100 (AUC, last in 2021); VOT2018 (Accuracy, last in 2021).
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2016 · Hedged Deep Tracking (CVPR 2016 · 741 citations)
- 2017 · Hierarchical Attentive Recurrent Tracking (NeurIPS 2017 · 51 citations)
- 2018 · High Performance Visual Tracking With Siamese Region Proposal Network (CVPR 2018 · 3,032 citations)
- 2019 · Fast Online Object Tracking and Segmentation: A Unifying Approach (CVPR 2019 · 1,541 citations)
- 2020 · Siamese Box Adaptive Network for Visual Tracking (CVPR 2020 · 1,015 citations)
- 2021 · Graph Attention Tracking (CVPR 2021 · 495 citations)
- 2022 · Correlation-Aware Deep Tracking (CVPR 2022 · 199 citations)
- 2023 · SeqTrack: Sequence to Sequence Learning for Visual Object Tracking (CVPR 2023 · 462 citations)
- 2024 · Learning Random Noise Salient Feature Fusion Siamese Network for Low-Resolution Object Tracking (Student Abstract) (AAAI 2024 · 3 citations)
- 2025 · Two-stream Beats One-stream: Asymmetric Siamese Network for Efficient Visual Tracking (AAAI 2025 · 25 citations)
Earliest papers here
- Hedged Deep Tracking (CVPR 2016 · 741 citations)
- Learning feed-forward one-shot learners (NeurIPS 2016 · 236 citations)
- Visual Tracking Using Attention-Modulated Disintegration and Integration (CVPR 2016 · 203 citations)
Benchmark records
The best number each year that beat every earlier one, among papers of this approach reporting their own method. Numbers are extracted by AI from each paper's text; check the paper before citing.
VOT2018 · EAO (higher is better)
8 papers, 2019–2021 · not reported since 2021
| Year | Best so far | Set by |
|---|---|---|
| 2019 | 0.38 | Fast Online Object Tracking and Segmentation: A Unifying Approach (CVPR 2019 · 1,541 citations) |
| 2020 | 0.481 | Discriminative and Robust Online Learning for Siamese Visual Tracking (AAAI 2020 · 63 citations) |
VOT2016 · EAO (higher is better)
7 papers, 2018–2020 · not reported since 2020
| Year | Best so far | Set by |
|---|---|---|
| 2018 | 0.344 | High Performance Visual Tracking With Siamese Region Proposal Network (CVPR 2018 · 3,032 citations) |
| 2019 | 0.363 | Siamese Cascaded Region Proposal Networks for Real-Time Visual Tracking (CVPR 2019 · 490 citations) |
| 2020 | 0.537 | Deformable Siamese Attention Networks for Visual Object Tracking (CVPR 2020 · 475 citations) |
VOT2015 · EAO (higher is better)
5 papers, 2018–2019 · not reported since 2019
| Year | Best so far | Set by |
|---|---|---|
| 2018 | 0.358 | High Performance Visual Tracking With Siamese Region Proposal Network (CVPR 2018 · 3,032 citations) |
OTB100 · AUC (higher is better)
4 papers, 2018–2021 · not reported since 2021
| Year | Best so far | Set by |
|---|---|---|
| 2018 | 0.657 | A Twofold Siamese Network for Real-Time Object Tracking (CVPR 2018 · 666 citations) |
| 2020 | 0.696 | Siamese Box Adaptive Network for Visual Tracking (CVPR 2020 · 1,015 citations) |
| 2021 | 0.713 | CapsuleRRT: Relationships-Aware Regression Tracking via Capsules (CVPR 2021 · 21 citations) |
TrackingNet · AUC (higher is better)
4 papers, 2021–2025 · one report since 2024, not better
| Year | Best so far | Set by |
|---|---|---|
| 2021 | 81.94 | Tracking Without Re-recognition in Humans and Machines (NeurIPS 2021 · 6 citations) |
| 2023 | 85.5 | SeqTrack: Sequence to Sequence Learning for Visual Object Tracking (CVPR 2023 · 462 citations) |
VOT2018 · Accuracy (higher is better)
4 papers, 2019–2021 · not reported since 2021
| Year | Best so far | Set by |
|---|---|---|
| 2019 | 0.609 | Fast Online Object Tracking and Segmentation: A Unifying Approach (CVPR 2019 · 1,541 citations) |
| 2020 | 0.63 | Deformable Siamese Attention Networks for Visual Object Tracking (CVPR 2020 · 475 citations) |
What has been tried
34 papers here report 342 measured changes (a component added, removed or swapped, compared with a baseline): helped 76%, no clear effect 4%, hurt 18%, mixed 2%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Add ETM | GOT-10k · AO | 60.4 -> 62.9 | Helped | Two-stream Beats One-stream: Asymmetric Siamese Network for Efficient Visual Tracking (AAAI 2025 · 25 citations) |
Baseline numbers that disagree
The same baseline on the same benchmark and metric, reported more than 5% apart by different papers here. Training budgets, data splits and reimplementations differ, so state which number you compare with.
| Baseline | Benchmark | Metric | Reported | Papers |
|---|---|---|---|---|
| SiamRPN | VOT-2016 | EAO | 0.344 – 0.464 | 2 papers, e.g. CVPR 2019 and CVPR 2020 |
| TransT | GOT-10k | SR0.75 | 60.9 – 68.2 | 2 papers, e.g. AAAI 2023 and NeurIPS 2021 |
| TransT | GOT-10k | AO | 67.1 – 72.3 | 2 papers, e.g. AAAI 2023 and NeurIPS 2021 |
| TransT | GOT-10k | SR0.5 | 76.8 – 82.4 | 2 papers, e.g. AAAI 2023 and NeurIPS 2021 |
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 5 papers; from OpenAlex and the papers' affiliations).
- Dalian University of Technology · 2 papers
Use it
Download BibTeX of 22 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Foundation-model and correlation-filter tracking (66 papers)
- Transformer-based video object tracking (74 papers)
Elsewhere in the topic
- Association cues for online MOT (60 papers)
- Memory and correlation-based MOT (52 papers)
- Unified transformer tracking (25 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).