Research map / Graph learning / Knowledge and temporal graphs
GPS-Net: Graph Property Sensing Network for Scene Graph Generation
Xin Lin, Changxing Ding, Jinquan Zeng, Dacheng Tao
Check 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 |
|---|---|---|---|
| Use GPS-Net rather than RelDN | SGG on VG · VG SGDET | R@100: 32.7 -> 33.2 | Improved |
| Use GPS-Net rather than VCTREE-HL | SGG on VG · VG SGCLS | R@100: 38.8 -> 42.3 | Improved |
| Use GPS-Net rather than RelDN | SGG on OI · OI | score_wtd: 44.61 -> 47.03 | Improved |
| Add DMP to the baseline (ablation) | GPS-Net baseline · VG SGCLS | R@100: 36.3 -> 39.3 | Improved |
| Add NPS-loss to the baseline (ablation) | GPS-Net baseline · VG SGCLS | R@100: 36.3 -> 37.1 | Improved |
| Add ARM to the baseline (ablation) | GPS-Net baseline · VG PREDCLS | R@100: 67.3 -> 68.5 | Improved |
| Use DMP instead of GCMP or S-GCMP (ablation) | GPS-Net message passing · VG SGCLS | R@100: 37.9 / 38.4 -> 40.1 | Improved |
| Add direction stacking in DMP (ablation) | GPS-Net · VG SGCLS | R@100: 40.1 -> 39.6 | Hurt |
| Replace focal loss with NPS-loss (ablation) | GPS-Net object classification · VG SGCLS | R@100: 39.8 -> 40.1 | Improved |
| Set NPS-loss controlling factor μ to 4 rather than 3 or 5 (ablation) | GPS-Net · VG SGCLS | R@100: 39.9 -> 40.1 | Improved |
Later papers that compared with this method
7 later accepted papers report GPSnet as a baseline in their results tables; 7 of them report a better number on most of the results they share with it.
- Panoptic Scene Graph Generation with Semantics-Prototype Learning (AAAI 2024 · better on 1 of 1 shared result)
- Constrained Structure Learning for Scene Graph Generation (TPAMI 2023 · better on 6 of 6 shared results)
- RelTR: Relation Transformer for Scene Graph Generation (TPAMI 2023 · better on 2 of 2 shared results)
- Neural Belief Propagation for Scene Graph Generation (TPAMI 2023 · better on 1 of 1 shared result)
- Panoptic Scene Graph Generation (ECCV 2022 · better on 4 of 4 shared results)
- RU-Net: Regularized Unrolling Network for Scene Graph Generation (CVPR 2022 · better on 1 of 1 shared result)
- Fine-Grained Scene Graph Generation with Data Transfer (ECCV 2022 · better on 1 of 1 shared result)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Scene Graph Generation Strategy with Co-occurrence Knowledge and Learnable Term Frequency (ICML 2024 · similarity 0.89)
- RU-Net: Regularized Unrolling Network for Scene Graph Generation (CVPR 2022 · similarity 0.88)
- A Comprehensive Survey of Scene Graphs: Generation and Application (TPAMI 2021 · similarity 0.88)
- Scene Graph-Grounded Image Generation (AAAI 2025 · similarity 0.87)
- Adaptive Visual Scene Understanding: Incremental Scene Graph Generation (NeurIPS 2024 · similarity 0.86)
- DSGG: Dense Relation Transformer for an End-to-end Scene Graph Generation (CVPR 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.
