Research map / Graph learning / Knowledge and temporal graphs
Multi-Granularity Sparse Relationship Matrix Prediction Network for End-to-End Scene Graph Generation
lei wang, Zejian Yuan, Badong Chen
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 semantic and positional information to Mg-RMPN input alongside visual features (ablation) | Mg-RMPN · VG150 SGDet | F@20/50/100: 13.4/18.2/21.7 -> 14.1/19.3/22.8 | Improved |
| use three gated pooling modules (multi-granularity negative filtering) instead of fewer (ablation) | Mg-RMPN · VG150 SGDet | F@20/50/100: full 14.1/19.3/22.8 F@20/50/100; GPM-1 or GPM-2 alone are very poor | Improved |
| replace Pair-Net's visual-only pair selection with the multi-granularity sparse relationship matrix | end-to-end SGG · VG150 SGDet | F@50: 16.6 -> 19.3 | Improved |
| use the proposed Mg-RMPN with DETR detector to remove detector influence | Mg-RMPN · VG150 SGDet | F@50: 18.0 vs Pair-Net 16.6 | Improved |
| rebalancing techniques in two-stage SGG (e.g. SHA+GCL) | two-stage SGG · VG150 SGDet | R@50: R@50 14.9 for SHA+GCL | Hurt |
| Mg-RMPN vs SHA+GCL rebalancing on GQA200 | Mg-RMPN · GQA200 SGDet | F@100: 18.5 vs 18.9 (SHA+GCL) | Mixed |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- RelTR — RelTR: Relation Transformer for Scene Graph Generation (TPAMI 2023) · better on 1 of 1 shared result
- SGTR — SGTR: End-to-End Scene Graph Generation With Transformer (CVPR 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
- DSGG: Dense Relation Transformer for an End-to-end Scene Graph Generation (CVPR 2024 · similarity 0.90)
- Leveraging Predicate and Triplet Learning for Scene Graph Generation (CVPR 2024 · similarity 0.90)
- EGTR: Extracting Graph from Transformer for Scene Graph Generation (CVPR 2024 · similarity 0.88)
- Structured Sparse R-CNN for Direct Scene Graph Generation (CVPR 2022 · similarity 0.88)
- Factorizable Net: An Efficient Subgraph-based Framework for Scene Graph Generation (ECCV 2018 · similarity 0.88)
- Hydra-SGG: Hybrid Relation Assignment for One-stage Scene Graph Generation (ICLR 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.
