Research map / Detection and segmentation / Tracking
Deformable Siamese Attention Networks for Visual Object Tracking
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 |
|---|---|---|---|
| add Siamese attention and region refinement | SiamRPN++ · VOT2016 | EAO: 0.464 -> 0.537 | Helped |
| add mask learning (ablation) | SiamRPN++ · VOT2016 | EAO: 0.464 -> 0.477 | Helped |
| add region refinement, including mask learning (ablation) | SiamRPN++ · VOT2016 | EAO: 0.464 -> 0.486 | Helped |
| add self-attention to region refinement (ablation) | SiamRPN++ · VOT2016 | EAO: +4.7% | Helped |
| add cross-attention to region refinement (ablation) | SiamRPN++ · VOT2016 | EAO: +4.9% | Helped |
| use self-attention and cross-attention jointly (ablation) | SiamAttn · VOT2016 | EAO: +7.3% | Helped |
| add deformable convolution (ablation) | SiamAttn · VOT2016 | EAO: 0.531 -> 0.537 | Helped |
| add deformable pooling (ablation) | SiamAttn · VOT2016 | EAO: 0.520 -> 0.537 | Helped |
| replace both deformable operations with regular operations (ablation) | SiamAttn · VOT2016 | EAO: 0.537 -> 0.516 | Hurt |
| train on COCO, YouTube-VOS, LaSOT, and TrackingNet instead of the alternative set (ablation) | SiamAttn · VOT2016 | EAO: 0.525 -> 0.537 | Helped |
| use SiamAttn instead of SiamRPN++ | visual object tracking · VOT2018 | EAO: 0.415 -> 0.470 | Helped |
| use SiamAttn instead of DiMP-50 | visual object tracking · TrackingNet | success: 74.0 -> 75.2 | Helped |
Closest papers by meaning
- Graph Attention Tracking (CVPR 2021 · similarity 0.92)
- Deeper and Wider Siamese Networks for Real-Time Visual Tracking (CVPR 2019 · similarity 0.92)
- SiamRPN++: Evolution of Siamese Visual Tracking With Very Deep Networks (CVPR 2019 · similarity 0.91)
- A Twofold Siamese Network for Real-Time Object Tracking (CVPR 2018 · similarity 0.91)
- Learning Attentions: Residual Attentional Siamese Network for High Performance Online Visual Tracking (CVPR 2018 · similarity 0.91)
- Discriminative and Robust Online Learning for Siamese Visual Tracking (AAAI 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.