Research map / Video / Video segmentation and tracking
Spatial-Temporal Transformer for Dynamic Scene Graph Generation
Yuren Cong, Wentong Liao, Hanno Ackermann, Bodo Rosenhahn, Michael Ying Yang
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 temporal decoder to spatial encoder (ablation) | STTran · Action Genome PredCLS (With Constraint) | R@20: 69.6 -> 71.3 | Improved |
| add spatial encoder to temporal decoder (ablation) | STTran · Action Genome PredCLS (With Constraint) | R@20: 71.0 -> 71.3 | Improved |
| add learned frame encoding (ablation) | STTran · Action Genome PredCLS (With Constraint) | R@20: 71.3 -> 71.8 | Improved |
| add learned frame encoding instead of sinusoidal (ablation) | STTran · Action Genome PredCLS (With Constraint) | R@20: 71.3 -> 71.8 | Improved |
| add temporal decoder (SGDET, With Constraint) (ablation) | STTran · Action Genome SGDET (With Constraint) | R@20: 32.9 -> 33.8 | Improved |
| add learned frame encoding (ablation) | STTran · Action Genome PredCLS (Semi Constraint) | R@20: 82.7 -> 83.1 | Improved |
| add LSTM over relationship features (ablation) | GPS-Net · Action Genome PredCLS (With Constraint) | R@20: 69.9 -> 70.4 | Improved |
| add LSTM over relationship features (ablation) | Motif Freq · Action Genome PredCLS (With Constraint) | R@20: 65.1 -> 65.2 | No clear effect |
| add LSTM over relationship features (ablation) | MSDN · Action Genome PredCLS (With Constraint) | R@20: 68.5 -> 68.8 | Improved |
| add LSTM over relationship features (ablation) | RelDN · Action Genome PredCLS (With Constraint) | R@20: 69.5 -> 69.7 | Improved |
| train on clean videos instead of 1/3 reversed (ablation) | STTran · Action Genome PredCLS (With Constraint) | R@20: 71.0 -> 71.8 | Improved |
| train on clean videos instead of 1/3 shuffled (ablation) | STTran · Action Genome PredCLS (With Constraint) | R@20: 70.6 -> 71.8 | Improved |
Closest papers by meaning
- OED: Towards One-stage End-to-End Dynamic Scene Graph Generation (CVPR 2024 · similarity 0.88)
- A Closer Look at Dynamic Scene Graph Generation in the Era of Multimodal Large Language Models (NeurIPS 2026 · similarity 0.88)
- Towards Unbiased and Robust Spatio-Temporal Scene Graph Generation and Anticipation (CVPR 2025 · similarity 0.87)
- TD²-Net: Toward Denoising and Debiasing for Video Scene Graph Generation (AAAI 2024 · similarity 0.87)
- DSGG: Dense Relation Transformer for an End-to-end Scene Graph Generation (CVPR 2024 · similarity 0.86)
- DiffVsgg: Diffusion-Driven Online Video Scene Graph Generation (CVPR 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.
