Research map / Detection and segmentation / Tracking
Siam R-CNN: Visual Tracking by Re-Detection
Check 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 |
|---|---|---|---|
| Use video hard example mining for re-detection training (ablation) | Siam R-CNN · OTB2015, LaSOT, LTB35 | AUC (OTB2015, LaSOT); F (LTB35): 68.4 -> 70.1; 63.2 -> 64.8; 66.5 -> 66.8 | Helped |
| Use TDPA instead of selecting the highest-scoring re-detection per frame (ablation) | Siam R-CNN · OTB2015, LaSOT, LTB35 | AUC (OTB2015, LaSOT); F (LTB35): 63.8 -> 70.1; 62.9 -> 64.8; 65.5 -> 66.8 | Helped |
| Use TDPA instead of the short-term tracking algorithm (ablation) | Siam R-CNN · OTB2015, LaSOT, LTB35 | AUC (OTB2015, LaSOT); F (LTB35): 67.2 -> 70.1; 55.7 -> 64.8; 57.2 -> 66.8 | Helped |
| Use a ResNet-50 rather than ResNet-101 backbone (ablation) | Siam R-CNN · OTB2015, LaSOT, LTB35 | AUC (OTB2015, LaSOT); F (LTB35): 70.1 -> 68.0; 64.8 -> 62.3; 66.8 -> 64.4 | Hurt |
| Use a two-stage full-image re-detection tracker with TDPA | Siam R-CNN · NfS | Success: 62.0 -> 63.9 | Helped |
| Use a two-stage full-image re-detection tracker with TDPA | Siam R-CNN · TrackingNet test | Success: 74.0 -> 81.2 | Helped |
| Use a two-stage full-image re-detection tracker with TDPA | Siam R-CNN · LaSOT test | Success rate: +8 percentage points | Helped |
| Use a two-stage full-image re-detection tracker with TDPA | Siam R-CNN · OxUvA test | MaxGM: 62.2 -> 72.3 | Helped |
| Add Box2Seg mask density filtering to rotated box estimation (ablation) | Siam R-CNN short-term · VOT2018 | EAO: 0.409 -> 0.423 | Helped |
Closest papers by meaning
- Deeper and Wider Siamese Networks for Real-Time Visual Tracking (CVPR 2019 · similarity 0.91)
- High Performance Visual Tracking With Siamese Region Proposal Network (CVPR 2018 · similarity 0.91)
- SiamCAR: Siamese Fully Convolutional Classification and Regression for Visual Tracking (CVPR 2020 · similarity 0.90)
- Discriminative and Robust Online Learning for Siamese Visual Tracking (AAAI 2020 · similarity 0.90)
- SiamFC++: Towards Robust and Accurate Visual Tracking with Target Estimation Guidelines (AAAI 2020 · similarity 0.90)
- A Twofold Siamese Network for Real-Time Object Tracking (CVPR 2018 · similarity 0.90)
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.
