Research map / Video / Video segmentation and tracking
Spatial Feature Calibration and Temporal Fusion for Effective One-Stage Video Instance Segmentation
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 feature calibration for anchors using larger aspect-ratio-specific kernels (ablation) | Yolact · YouTube-VIS valid | AP: 28.9 -> 30.8 | Helped |
| Use FCA with smaller aspect-ratio-specific kernels (ablation) | Yolact · YouTube-VIS valid | AP: +1% | Helped |
| Add adaptive bounding-box feature calibration (ablation) | Baseline · YouTube-VIS valid | AP: +2.3% | Helped |
| Add aligned bounding-box feature calibration (ablation) | Baseline · YouTube-VIS valid | AP: +2.4% | Helped |
| Add temporal fusion (ablation) | Baseline · YouTube-VIS valid | AP: +5.2% | Helped |
| Add adaptive FCB after FCA (ablation) | Yolact + FCA · YouTube-VIS valid | AP: 30.8 -> 34.4 | Helped |
| Add aligned FCB after FCA (ablation) | Yolact + FCA · YouTube-VIS valid | AP: 30.8 -> 33.1 | Helped |
| Add temporal fusion to FCA plus adaptive FCB (ablation) | Yolact + FCA + FCB (ada) · YouTube-VIS valid | AP: 34.4 -> 36.8 | Helped |
| Add temporal fusion to FCA plus aligned FCB (ablation) | Yolact + FCA + FCB (ali) · YouTube-VIS valid | AP: 33.1 -> 36.3 | Helped |
| Use adaptive rather than aligned FCB features (ablation) | STMask · YouTube-VIS valid | AP: 36.3 -> 36.8 | Mixed |
| Calibrate classification, mask, and tracking branches together with adaptive features (ablation) | Baseline + FCA · YouTube-VIS valid | AP: 30.8 -> 33.7 | Helped |
| Calibrate only classification and mask branches with adaptive features (ablation) | Baseline + FCA · YouTube-VIS valid | AP: 30.8 -> 34.4 | Helped |
Closest papers by meaning
- Hybrid Instance-Aware Temporal Fusion for Online Video Instance Segmentation (AAAI 2022 · similarity 0.90)
- Mask-Free Video Instance Segmentation (CVPR 2023 · similarity 0.89)
- S2D: Sparse-To-Dense Keymask Distillation for Unsupervised Video Instance Segmentation (NeurIPS 2026 · similarity 0.88)
- SG-Net: Spatial Granularity Network for One-Stage Video Instance Segmentation (CVPR 2021 · similarity 0.88)
- Robust and Consistent Online Video Instance Segmentation via Instance Mask Propagation (AAAI 2025 · similarity 0.87)
- VideoCutLER: Surprisingly Simple Unsupervised Video Instance 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.