Research map / Detection and segmentation / Tracking
Alpha-Refine: Boosting Tracking Performance by Precise Bounding Box Estimation
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 Alpha-Refine to DiMPsuper | DiMPsuper · LaSOT test | AUC: 63.7 -> 65.3 | Helped |
| Add Alpha-Refine to DiMPsuper | DiMPsuper · TrackingNet test | AUC: 77.6 -> 80.5 | Helped |
| Add Alpha-Refine to DiMPsuper | DiMPsuper · GOT-10K test | AO: 67.2 -> 70.1 | Helped |
| Add Alpha-Refine to DiMPsuper | DiMPsuper · VOT2020 real-time | EAO: 0.311 -> 0.478 | Helped |
| Add an auxiliary mask head to the corner-head refinement module (ablation) | SiamRPNpp+AR · LaSOT test | AUC: 54.6 -> 55.9 | Helped |
| Replace the corner head with an RPN-style head while retaining the mask head (ablation) | SiamRPNpp+AR · LaSOT test | AUC: 55.9 -> 53.7 | Hurt |
| Replace the corner head with an RCNN-style head while retaining the mask head (ablation) | SiamRPNpp+AR · LaSOT test | AUC: 55.9 -> 51.6 | Hurt |
| Replace the corner head with an RPN-style head and omit the mask head (ablation) | SiamRPNpp+AR · LaSOT test | AUC: 55.9 -> 50.2 | Hurt |
| Replace the corner head with an RCNN-style head and omit the mask head (ablation) | SiamRPNpp+AR · LaSOT test | AUC: 55.9 -> 48.9 | Hurt |
| Replace pixel-wise correlation with depth-wise correlation (ablation) | SiamRPNpp+AR · LaSOT test | AUC: 55.9 -> 54.8 | Hurt |
| Replace pixel-wise correlation with naive correlation (ablation) | SiamRPNpp+AR · LaSOT test | AUC: 55.9 -> 53.1 | Hurt |
| Replace ResNet-34 with ResNet-50 (ablation) | SiamRPNpp+AR · LaSOT test | AUC: 55.9 -> 56.2 | Helped |
Closest papers by meaning
- ATOM: Accurate Tracking by Overlap Maximization (CVPR 2019 · similarity 0.90)
- Learning Patch-Based Dynamic Graph for Visual Tracking (AAAI 2017 · similarity 0.86)
- Correlation-Guided Attention for Corner Detection Based Visual Tracking (CVPR 2020 · similarity 0.86)
- DeTrack: In-model Latent Denoising Learning for Visual Object Tracking (NeurIPS 2024 · similarity 0.85)
- Focusing on Tracks for Online Multi-Object Tracking (CVPR 2025 · similarity 0.84)
- DiffuBox: Refining 3D Object Detection with Point Diffusion (NeurIPS 2024 · similarity 0.84)
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.