Research map / Video / Video-language understanding
SOC: Semantic-Assisted Object Cluster for Referring Video Object Segmentation
Read the paperCheck 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 video-level object cluster (VOC) (ablation) | SOC · Ref-YouTube-VOS | J&F: 55.6 -> 58.0 | Helped |
| add visual-linguistic contrastive loss (ablation) | SOC · Ref-YouTube-VOS | J&F: 55.6 -> 57.1 | Helped |
| add both VOC and contrastive loss (ablation) | SOC · Ref-YouTube-VOS | J&F: 55.6 -> 59.2 | Helped |
| add language-to-vision fusion (ablation) | SOC · Ref-YouTube-VOS | J&F: 39.8 -> 58.7 | Helped |
| add vision-to-language fusion on top of L2V (ablation) | SOC · Ref-YouTube-VOS | J&F: 58.7 -> 59.2 | Helped |
| object query number 10 to 20 (ablation) | SOC · Ref-YouTube-VOS | J&F: 58.0 -> 59.2 | Helped |
| increase object queries from 20 to 25 (ablation) | SOC · Ref-YouTube-VOS | J&F: 59.2 -> 58.5 | Hurt |
| increase training frames from 3 to 8 (ablation) | SOC · Ref-YouTube-VOS | J&F: 57.9 -> 59.2 | Helped |
| increase training frames from 8 to 10 (ablation) | SOC · Ref-YouTube-VOS | J&F: 59.2 -> 59.1 | No clear effect |
| add VOC encoder and decoder both vs encoder only (ablation) | SOC · Ref-YouTube-VOS | J&F: 58.3 -> 59.2 | Helped |
| add VOC encoder and decoder both vs decoder only (ablation) | SOC · Ref-YouTube-VOS | J&F: 58.1 -> 59.2 | Helped |
| SOC vs ReferFormer (Swin-T, scratch) (ablation) | SOC vs ReferFormer · Ref-YouTube-VOS | J&F: 56.0 -> 59.2 | Helped |
Closest papers by meaning
- Semantic and Sequential Alignment for Referring Video Object Segmentation (CVPR 2025 · similarity 0.94)
- LoSh: Long-Short Text Joint Prediction Network for Referring Video Object Segmentation (CVPR 2024 · similarity 0.91)
- SAMWISE: Infusing Wisdom in SAM2 for Text-Driven Video Segmentation (CVPR 2025 · similarity 0.91)
- Referred by Multi-Modality: A Unified Temporal Transformer for Video Object Segmentation (AAAI 2024 · similarity 0.89)
- Learning Spatial-Semantic Features for Robust Video Object Segmentation (ICLR 2025 · similarity 0.88)
- Exploiting Temporal State Space Sharing for Video Semantic Segmentation (CVPR 2025 · 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.