Research map / Detection and segmentation / Tracking
Triplet Loss in Siamese Network for Object Tracking
Xingping Dong, Jianbing Shen
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 |
|---|---|---|---|
| replace pairwise logistic loss with triplet loss | SiamFC · VOT-2017 real-time | EAO: 0.1966 -> 0.2125 | Improved |
| replace pairwise logistic loss with triplet loss | SiamImp · VOT-2017 real-time | EAO: 0.1728 -> 0.1833 | Improved |
| replace pairwise logistic loss with triplet loss | CFnet2 · VOT-2017 real-time | EAO: 0.0963 -> 0.1080 | Improved |
| replace logistic loss with triplet loss | SiamImp · OTB-50 | success AUC: 3.9% improvement | Improved |
| replace logistic loss with triplet loss | SiamFC · OTB-100 low resolution subset | success AUC: 0.619 -> 0.615 | Hurt |
| train 10 more epochs with original logistic loss (-init variants) (ablation) | SiamFC, SiamImp · OTB-2013 | precision and AUC | Hurt |
| train 10 more epochs with triplet loss instead of logistic loss from same init (ablation) | SiamFC, CFnet2, SiamImp · OTB-2013 | precision and success AUC | Improved |
| swap logistic loss for triplet loss (tracking speed) | SiamFC · OTB-2013 | fps: 86.5 -> 86.3 | No clear effect |
| swap logistic loss for triplet loss (tracking speed) | SiamImp · OTB-2013 | fps: 55.4 -> 55.8 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- Staple — Staple: Complementary Learners for Real-Time Tracking (CVPR 2016) · better on 1 of 2 shared results
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
- Deeper and Wider Siamese Networks for Real-Time Visual Tracking (CVPR 2019 · similarity 0.90)
- SiamRPN++: Evolution of Siamese Visual Tracking With Very Deep Networks (CVPR 2019 · similarity 0.90)
- High Performance Visual Tracking With Siamese Region Proposal Network (CVPR 2018 · similarity 0.89)
- Siamese Instance Search for Tracking (CVPR 2016 · similarity 0.89)
- Deformable Siamese Attention Networks for Visual Object Tracking (CVPR 2020 · similarity 0.88)
- Learning Dynamic Siamese Network for Visual Object Tracking (ICCV 2017 · similarity 0.88)
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.
