Research map / Detection and segmentation / Tracking
GradNet: Gradient-Guided Network for Visual Object Tracking
Peixia Li, Boyu Chen, Wanli Ouyang, Dong Wang, Xiaoyun Yang, Huchuan Lu
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 template generalization training (one template across search regions from several videos) (ablation) | GradNet · OTB-2015 | precision: 0.823 -> 0.861 | Improved |
| add template generalization training (ablation) | GradNet · OTB-2015 | success (IOU): 0.615 -> 0.639 | Improved |
| add gradient-guided update branch on top of SiameseFC (full method vs baseline) | SiameseFC · OTB-2015 | precision: 0.771 -> 0.861 | Improved |
| add template generalization and gradient application (vs two unshared branches) (ablation) | GradNet · OTB-2015 | precision: 0.717 -> 0.861 | Improved |
| add online template update (ablation) | GradNet · OTB-2015 | precision: 0.775 -> 0.861 | Improved |
| add online template update (ablation) | GradNet · OTB-2015 | FPS: 85 -> 80 | Hurt |
| share parameters between sub-nets U1 and U2 (vs separate parameters) (ablation) | GradNet · OTB-2015 | precision: 0.833 -> 0.861 | Improved |
| share parameters between sub-nets U1 and U2 (ablation) | GradNet · OTB-2015 | success (IOU): 0.628 -> 0.639 | Improved |
| add gradient-guided template update to SiameseFC | SiameseFC · VOT2017 | EAO: 0.182 -> 0.247 | Improved |
| add gradient-guided template update to SiameseFC | SiameseFC · VOT2017 | robustness failures (R): 0.604 -> 0.375 | Improved |
| use GradNet versus SiamRPN trained on far more data | GradNet vs SiamRPN · VOT2017 | EAO: 0.247 vs 0.244 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Staple — Staple: Complementary Learners for Real-Time Tracking (CVPR 2016) · better on 1 of 1 shared result
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
- Joint Classification and Regression for Visual Tracking with Fully Convolutional Siamese Networks (IJCV 2022 · similarity 0.91)
- High Performance Visual Tracking With Siamese Region Proposal Network (CVPR 2018 · similarity 0.91)
- Learning the Model Update for Siamese Trackers (ICCV 2019 · similarity 0.90)
- Learning Dynamic Siamese Network for Visual Object Tracking (ICCV 2017 · similarity 0.90)
- Deeper and Wider Siamese Networks for Real-Time Visual Tracking (CVPR 2019 · similarity 0.90)
- SiamCAR: Siamese Fully Convolutional Classification and Regression for Visual Tracking (CVPR 2020 · 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.
