Research map / Video / Video segmentation and tracking
Robust Referring Video Object Segmentation with Cyclic Structural Consensus
Xiang Li, Jinglu Wang, Xiaohao Xu, Xiao Li, Bhiksha Raj, Yan Lu
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 object localizing module (OLM) (ablation) | R2-VOS baseline · Ref-Youtube-VOS (R-VOS) | J&F: 52.4 -> 55.5 | Improved |
| add cycle consistency constraint (ablation) | R2-VOS with OLM · Ref-Youtube-VOS (R-VOS) | J&F: 55.5 -> 56.9 | Improved |
| add text embedding fusion (ablation) | R2-VOS · Ref-Youtube-VOS (R-VOS) | J&F: 56.9 -> 57.3 | Improved |
| add OLM, consistency constraint and text fusion to baseline (ablation) | R2-VOS baseline · R2-Youtube-VOS | A: 53.2 -> 64.5 | Improved |
| use structural (distance + angle) consistency instead of point-wise (ablation) | R2-VOS · R2-Youtube-VOS | J&F: 52.4 -> 56.8 | Improved |
| combine structural distance and angle constraints (ablation) | R2-VOS · R2-Youtube-VOS | A: 60.0 (SA) / 59.3 (SD) -> 62.1 | Improved |
| use 5 instead of 1 instance query (ablation) | R2-VOS · R2-Youtube-VOS | J&F: 54.9 -> 57.2 | Improved |
| use 9 instead of 5 instance queries (ablation) | R2-VOS · R2-Youtube-VOS | J&F: 57.2 -> 57.0 | No clear effect |
| train R2-VOS instead of ReferFormer on R2-VOS task | R-VOS methods · R2-Youtube-VOS | J&F: 42.2 -> 57.2 | Improved |
| replace ReferFormer with R2-VOS on standard R-VOS | R-VOS · Ref-Youtube-VOS (ResNet50) | J&F: 55.6 -> 57.3 | Improved |
| R2-VOS speed vs ReferFormer | R-VOS · single P40 batchsize 1 | FPS: 22 -> 30 FPS | Improved |
Closest papers by meaning
- Language As Queries for Referring Video Object Segmentation (CVPR 2022 · similarity 0.91)
- SOC: Semantic-Assisted Object Cluster for Referring Video Object Segmentation (NeurIPS 2023 · similarity 0.90)
- Semantic and Sequential Alignment for Referring Video Object Segmentation (CVPR 2025 · similarity 0.89)
- LoSh: Long-Short Text Joint Prediction Network for Referring Video Object Segmentation (CVPR 2024 · similarity 0.89)
- Point-VOS: Pointing Up Video Object Segmentation (CVPR 2024 · similarity 0.87)
- Referred by Multi-Modality: A Unified Temporal Transformer for Video Object Segmentation (AAAI 2024 · similarity 0.87)
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.
