Research map / Video / Video segmentation and tracking
MiNQVIS: Mitigating Noisy Queries for Robust Online Video Instance Segmentation
Jie Qiao, Jianxu Chen, Xiaowei Xu
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 query-level gating (ablation) | CTVIS baseline · YouTube-VIS 2021 / OVIS | AP: 47.5 -> 49.8 / 34.7 -> 35.9 | Improved |
| add VPCL (ablation) | CTVIS baseline · YouTube-VIS 2021 / OVIS | AP: 47.5 -> 48.7 / 34.7 -> 36.3 | Improved |
| add dual-memory fusion (ablation) | CTVIS baseline · YouTube-VIS 2021 / OVIS | AP: 47.5 -> 48.1 / 34.7 -> 35.5 | Improved |
| combine gating and VPCL (ablation) | baseline · YouTube-VIS 2021 / OVIS | AP: 47.5 -> 51.4 / 34.7 -> 37.8 | Improved |
| add dual-memory fusion on top of gating + VPCL (ablation) | MiNQVIS · YouTube-VIS 2021 / OVIS | AP: 51.4 -> 51.7 / 37.8 -> 38.2 | Improved |
| apply gate only at training (ablation) | MiNQVIS · YouTube-VIS 2021 / OVIS | AP: 51.7 -> 39.6 / 38.2 -> 25.2 | Hurt |
| apply gate only at inference (ablation) | MiNQVIS · YouTube-VIS 2021 / OVIS | AP: 51.7 -> 48.3 / 38.2 -> 34.3 | Hurt |
| combine momentum and history-best memory (ablation) | MiNQVIS · YouTube-VIS 2021 / OVIS | AP: momentum 51.4 -> 51.7; OVIS 37.8 -> 38.2 | Improved |
| raise gate threshold gamma to 0.20 from 0.05-0.10 (ablation) | MiNQVIS · OVIS | AP: 38.2 -> 32.8 | Hurt |
| use MiNQVIS vs CAVIS under Swin-L on YouTube-VIS 2019 | online VIS · YouTube-VIS 2019 (Swin-L) | AP: 66.0 -> 66.2 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- CAVIS — CAVIS: Context-Aware Video Instance Segmentation (ICCV 2025) · better on 6 of 6 shared results
- CTVIS — CTVIS: Consistent Training for Online Video Instance Segmentation (ICCV 2023) · better on 6 of 6 shared results
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
- MinVIS: A Minimal Video Instance Segmentation Framework without Video-based Training (NeurIPS 2022 · similarity 0.93)
- In Defense of Online Models for Video Instance Segmentation (ECCV 2022 · similarity 0.92)
- CTVIS: Consistent Training for Online Video Instance Segmentation (ICCV 2023 · similarity 0.91)
- Minimizing Labeled, Maximizing Unlabeled: An Image-Driven Approach for Video Instance Segmentation (CVPR 2025 · similarity 0.91)
- SyncVIS: Synchronized Video Instance Segmentation (NeurIPS 2024 · similarity 0.90)
- Robust and Consistent Online Video Instance Segmentation via Instance Mask Propagation (AAAI 2025 · similarity 0.89)
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.
