Research map / Detection and segmentation / Open-set and few-shot detection
Deep Variation-Structured Reinforcement Learning for Visual Relationship and Attribute Detection
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 |
|---|---|---|---|
| Use VRL instead of Lu et al. | relationship phrase detection on VRD · VRD | Phr. R@100: 17.03 -> 22.60 | Helped |
| Use VRL instead of Lu et al. | relationship detection on Visual Genome · Visual Genome | Rel. R@100: 7.96 -> 13.34 | Helped |
| Use VRL instead of Joint CNN+Trained RPN | attribute detection · Visual Genome | Attribute Recall@100: 9.77 -> 26.43 | Helped |
| Use variation-structured action sets instead of the full action space (ablation) | VRL · Visual Genome | Rel. R@100: 6.23 -> 13.34 | Helped |
| Add history phrase embeddings to the state (ablation) | VRL · Visual Genome | Rel. R@100: 9.05 -> 13.34 | Helped |
| Use semantic phrase embeddings instead of historical action vectors (ablation) | VRL · Visual Genome | Rel. R@100: 11.98 -> 13.34 | Helped |
| Use category-based object mining instead of directional actions (ablation) | VRL · Visual Genome | Rel. R@100: 10.66 -> 13.34 | Helped |
| Include ambiguous candidate object categories during object mining (ablation) | VRL · Visual Genome | Rel. R@100: 12.01 -> 13.34 | Helped |
| Replace fully connected layers with LSTM layers (ablation) | VRL · Visual Genome | Rel. R@100: 13.34 -> 13.86 | No clear effect |
| Use sequential RL mining instead of random-walk traversal (ablation) | relationship and attribute detection · Visual Genome | Rel. R@100: 3.67 -> 6.23 | Helped |
| Use VRL rather than Lu et al. for zero-shot relationship detection | VRD · VRD zero-shot | Rel. R@100: 3.28 -> 8.52 | Helped |
Closest papers by meaning
- Large-Scale Visual Relationship Understanding (AAAI 2019 · similarity 0.89)
- Detecting Visual Relationships With Deep Relational Networks (CVPR 2017 · similarity 0.88)
- OVOD-Agent: A Markov-Bandit Framework for Proactive Visual Reasoning and Self-Evolving Detection (CVPR 2026 · similarity 0.87)
- Object-Centric Representation Learning for Enhanced 3D Semantic Scene Graph Prediction (NeurIPS 2025 · similarity 0.87)
- Synthetic Visual Genome (CVPR 2025 · similarity 0.87)
- Multi-Modal Prompting for Open-Vocabulary Video Visual Relationship Detection (AAAI 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.