Research map / Graph learning / GNN architectures
A Survey on Graph Neural Networks and Graph Transformers in Computer Vision: A Task-Oriented Perspective
Chaoqi Chen, Yushuang Wu, Qiyuan Dai, Hong-Yu Zhou, Mutian Xu, Sibei Yang, Xiaoguang Han, Yizhou Yu
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 label-graph GCN over class labels | ResNet-101 · MS-COCO multi-label classification | mAP: 77.3 -> 83.0 | Improved |
| use TDRG over ML-GCN | ML-GCN · MS-COCO multi-label classification | mAP: 83.0 -> 86.0 | Improved |
| model spatial skeleton graph with ST-GCN instead of LSTM | Deep LSTM · NTU-RGB+D X-Sub skeleton action recognition | accuracy: 60.7 -> 81.5 | Improved |
| use multi-scale MS-G3D graph convolution over ST-GCN | ST-GCN · NTU-RGB+D X-Sub skeleton action recognition | accuracy: 81.5 -> 91.5 | Improved |
| use MS-G3D over ST-GCN on Kinetics Skeleton | ST-GCN · Kinetics Skeleton skeleton action recognition | Top-1 accuracy: 30.7 -> 38.0 | Improved |
| add object relation graph (OR2G) to NL + I3D | NL + I3D · Charades action recognition | mAP: 37.5 -> 44.9 | Improved |
| add graph-based relation reasoning (ReGAT) to Bottom-Up | Bottom-Up · VQA v2.0 (std) | accuracy: 65.67 -> 70.58 | Improved |
| add single-hop LCGN graph reasoning to Bottom-Up | Bottom-Up · GQA test | accuracy: 49.7 -> 56.1 | Improved |
Closest papers by meaning
- Can Classic GNNs Be Strong Baselines for Graph-level Tasks? Simple Architectures Meet Excellence (ICML 2025 · similarity 0.90)
- Learning Convolutional Neural Networks for Graphs (ICML 2016 · similarity 0.89)
- Dynamic Graph Message Passing Networks for Visual Recognition (TPAMI 2022 · similarity 0.89)
- EdgeNets: Edge Varying Graph Neural Networks (TPAMI 2021 · similarity 0.89)
- A Manifold Perspective on the Statistical Generalization of Graph Neural Networks (ICML 2025 · similarity 0.88)
- Classic GNNs are Strong Baselines: Reassessing GNNs for Node Classification (NeurIPS 2024 · 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.
