Research map / Video / Video-language understanding
Few-Shot Referring Relationships in Videos
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 |
|---|---|---|---|
| use the full few-shot relationship grounding method | few-shot referring relationships in videos · ImageNet-VidVRD | Ar_s-t: 9.6 -> 25.1 | Helped |
| use the full few-shot relationship grounding method | few-shot referring relationships in videos · VidOR | Ar_s-t: 8.2 -> 23.8 | Helped |
| add visual relationship similarity potential (ablation) | full model · ImageNet-VidVRD | Ar_s-t: 17.2 -> 25.1 | Helped |
| add visual relationship similarity potential (ablation) | full model · VidOR | Ar_s-t: 17.6 -> 23.8 | Helped |
| add GSA and LLA (ablation) | few-shot relationship grounding model · ImageNet-VidVRD | Ar_s-t: 22.8 -> 25.1 | Helped |
| add GSA and LLA (ablation) | few-shot relationship grounding model · VidOR | Ar_s-t: 21.9 -> 23.8 | Helped |
| add GSA (ablation) | few-shot relationship grounding model · ImageNet-VidVRD | Ar_s-t: 20.8 -> 22.8 | Helped |
| add LLA (ablation) | few-shot relationship grounding model · ImageNet-VidVRD | Ar_s-t: 20.8 -> 21.1 | Helped |
| replace concatenation with translational relationship embedding (ablation) | few-shot relationship grounding model · ImageNet-VidVRD | Ar_s-t: 18.7 -> 25.1 | Helped |
| replace greedy trajectory selection with belief propagation (ablation) | few-shot relationship grounding model · ImageNet-VidVRD | Ar_s-t: 22.8 -> 25.1 | Helped |
| increase support set size (ablation) | few-shot relationship grounding model · ImageNet-VidVRD | Ar_s-t | Helped |
| evaluate when subject or object is unseen during training (ablation) | few-shot relationship grounding model · ImageNet-VidVRD | Ar_s-t: 27.3 -> 18.6; 27.3 -> 17.2 | Hurt |
Closest papers by meaning
- Localize, Assemble, and Predicate: Contextual Object Proposal Embedding for Visual Relation Detection (AAAI 2020 · similarity 0.89)
- Referring Relationships (CVPR 2018 · similarity 0.89)
- Multi-Modal Prompting for Open-Vocabulary Video Visual Relationship Detection (AAAI 2024 · similarity 0.88)
- Zero-shot Referring Expression Comprehension via Structural Similarity Between Images and Captions (CVPR 2024 · similarity 0.88)
- Dense Video Object Captioning from Disjoint Supervision (ICLR 2025 · similarity 0.86)
- SAMA: Towards Multi-Turn Referential Grounded Video Chat with Large Language Models (NeurIPS 2025 · 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.