Research map / Vision-language and multimodal / Human and embodied modeling
SPADE: Spatial-Aware Denoising Network for Open-vocabulary Panoptic Scene Graph Generation with Long- and Local-range Context Reasoning
Xin Hu, Ke Qin, Guiduo Duan, Ming Li, Yuan-Fang Li, Tao He
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 inversion-guided calibration (fine-tune UNet) (ablation) | SPADE · PSG open-set OvR | R@50: 15.3 -> 26.7 | Improved |
| use DDIM inversion sampling instead of random noise (ablation) | SPADE calibration · PSG open-set OvR | R@50: 21.0 -> 26.7 | Improved |
| use LoRA instead of fine-tuning all UNet parameters (ablation) | SPADE calibration · PSG open-set OvR | R@50: 18.8 -> 26.7 | Improved |
| add LCNL (ablation) | Relation Graph Transformer · PSG close-set | R@50: 30.5 -> 35.6 | Improved |
| add LCNNL (ablation) | Relation Graph Transformer · PSG close-set | R@50: 35.6 -> 37.1 | Improved |
| add local correlation learning (LCL) (ablation) | Relation Graph Transformer · PSG close-set | R@50: 37.1 -> 40.3 | Improved |
| add auxiliary relation query construction loss L_rqc (ablation) | Relation Graph Transformer · PSG close-set | R@50: 40.3 -> 45.1 | Improved |
| add pooled features to diffusion features (ablation) | open-vocabulary module · PSG open-set OvR | R@50: 21.4 -> 26.7 | Improved |
| add diffusion features to pooled features (ablation) | open-vocabulary module · PSG open-set OvR | R@50: 12.5 -> 26.7 | Improved |
| use SPADE instead of OpenPSG on distant pairs | spatial predicate prediction · PSG DR | R@50: 37.1 -> 42.3 | Improved |
| use SPADE instead of OpenPSG on non-distant pairs | spatial predicate prediction · PSG NDR | R@50: 43.7 -> 46.5 | Improved |
| use SPADE instead of OpenPSG | open-set OvD+R · PSG | R@50: 11.4 -> 22.7 | Improved |
Closest papers by meaning
- DSGG: Dense Relation Transformer for an End-to-end Scene Graph Generation (CVPR 2024 · similarity 0.90)
- Panoptic Scene Graph Generation with Semantics-Prototype Learning (AAAI 2024 · similarity 0.89)
- From Pixels to Graphs: Open-Vocabulary Scene Graph Generation with Vision-Language Models (CVPR 2024 · similarity 0.88)
- Noise-Guided Predicate Representation Extraction and Diffusion-Enhanced Discretization for Scene Graph Generation (ICML 2025 · similarity 0.88)
- TD²-Net: Toward Denoising and Debiasing for Video Scene Graph Generation (AAAI 2024 · similarity 0.88)
- Synthetic Visual Genome (CVPR 2025 · similarity 0.88)
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.
