AI Paper Atlas

Research map / Graph learning / Knowledge and temporal graphs

Graph R-CNN for Scene Graph Generation

Jianwei Yang, Jiasen Lu, Stefan Lee, Dhruv Batra, Devi Parikh

ECCV 2018 · Graph learning › Knowledge and temporal graphs · approach: Scene graph generation · 919 citations (OpenAlex)

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 changedWhereMeasuredEffect
add the relation proposal network (RePN) (ablation)Graph R-CNN baseline · Visual GenomeSGGen R@50: 6.1 -> 8.7Improved
add RePN (ablation)Graph R-CNN baseline · Visual GenomeSGGen+ R@100: 27.9 -> 34.8Improved
add RePN (ablation)Graph R-CNN baseline · Visual Genomedetection mAP@0.5: 20.4 -> 23.6Improved
add RePN (ablation)Graph R-CNN baseline · Visual Genome (ground-truth boxes)PredCls R@50: 33.5 -> 34.5No clear effect
add plain GCN on top of RePN (ablation)Graph R-CNN with RePN · Visual GenomePredCls R@50: 34.5 -> 52.3Improved
add plain GCN on top of RePN (ablation)Graph R-CNN with RePN · Visual Genomedetection mAP@0.5: 23.6 -> 23.4No clear effect
use attentional GCN instead of plain GCN (ablation)Graph R-CNN with RePN · Visual GenomeSGGen R@50: 10.8 -> 11.4Improved
use attentional GCN instead of plain GCN (ablation)Graph R-CNN with RePN · Visual GenomePredCls R@50: 52.3 -> 54.2Improved
use attentional GCN instead of plain GCN (ablation)Graph R-CNN with RePN · Visual Genomedetection mAP@0.5: 23.4 -> 23.0Hurt
use Graph R-CNN instead of reimplemented IMPscene graph generation · Visual GenomeSGGen R@50: 6.4 -> 11.4Improved
use Graph R-CNN instead of frequency baseline NM-Freqscene graph generation · Visual GenomeSGGen+ R@100: 27.8 -> 35.9Improved
use SGGen+ instead of SGGen when 20% of objects with relationships are perturbedevaluation metric · perturbed ground-truth graphsrecall: 54.1 vs 84.3Mixed

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy