Research map / Detection and segmentation / Open-set and few-shot detection
Integrating Boxes and Masks: A Multi-Object Framework for Unified Visual Tracking and Segmentation
Yuanyou Xu, Zongxin Yang, Yi Yang
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 auxiliary mask reconstruction during training (ablation) | MITS UIDM · LaSOT | AUC: 71.5 -> 72.0 | Improved |
| use dual cross attention instead of single cross attention in BIDR (ablation) | MITS BIDR · LaSOT | AUC: 70.9 -> 72.0 | Improved |
| use pinpoint head instead of corner head (ablation) | MITS box branch · LaSOT | AUC: 69.1 -> 72.0 | Improved |
| explicit pinpoint probability maps instead of implicit localization (ablation) | MITS box branch · LaSOT | AUC: 71.1 -> 72.0 | Improved |
| train box branch jointly with mask branch (ablation) | MITS mask branch · LaSOT | AUC: 71.5 -> 72.3 | Improved |
| use transformer localization instead of FPN (ablation) | MITS box head · LaSOT | AUC: 70.6 -> 71.1 | Improved |
| use GPM instead of LSTT propagation module (ablation) | MITS · LaSOT | AUC: 69.3 -> 72.0 | Improved |
| use pseudo-masks for VOT training data (ablation) | MITS · LaSOT | AUC: 70.7 -> 72.0 | Improved |
| lighter MobileNetV2 backbone with 2 propagation layers (ablation) | MITS · LaSOT | AUC: 72.0 -> 67.8 AUC, 20 -> 27 FPS | Mixed |
| evaluate mask decoder output instead of box predictor | MITS · LaSOT | AUC: 72.0 -> 72.3 AUC; P 78.5 -> 78.2 | Mixed |
| use MITS with box initialization instead of mask initialization | MITS · YouTube-VOS 2019 val | G: 85.9 -> 81.8 | Hurt |
Closest papers by meaning
- Generating Masks From Boxes by Mining Spatio-Temporal Consistencies in Videos (ICCV 2021 · similarity 0.89)
- MHP-VOS: Multiple Hypotheses Propagation for Video Object Segmentation (CVPR 2019 · similarity 0.89)
- Domain Generalization for Multiple Video Object Segmentation and Tracking Using Transformers and Smart Memory (IJCV 2026 · similarity 0.89)
- Multi-Granularity Video Object Segmentation (AAAI 2025 · similarity 0.87)
- OVT-B: A New Large-Scale Benchmark for Open-Vocabulary Multi-Object Tracking (NeurIPS 2024 · similarity 0.86)
- Point-VOS: Pointing Up Video Object Segmentation (CVPR 2024 · similarity 0.86)
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.
