Graph Neural Networks

Also known as GNN

Graph Neural Networks (GNNs) study how to learn from data whose entities and relationships form a graph, supporting tasks such as node classification, link prediction, and graph or molecular property prediction. Common approaches include message-passing networks such as GCN, GraphSAGE, and GIN, attention-based Graph Transformers, and spectral, geometric, or higher-order methods that encode richer structure. They are used on benchmarks such as Cora and ogbn-arxiv, as well as in molecular modeling, recommender systems, scientific simulation, and network control. Key open challenges include handling heterophilic graphs, preserving useful information in deep networks without oversmoothing or oversquashing, and achieving reliable generalization, robustness, and efficiency on large or shifting graphs.

Latest papers

751 papers73 papers in the last four weeks, down 13% on the four weeks before. 0.5% of all new papers.

MUG: Multi-human Graph Network for 3D Mesh Reconstruction from 2D Pose
May 25, 2022Chenyan Wu, Yandong Li, Xianfeng Tang +23D Human Reconstruction Graph Neural Networks
The Pennsylvania State University · Google · Amazon

May 25, 2022J/K move · Enter open · S save

MUG: Multi-human Graph Network for 3D Mesh Reconstruction from 2D Pose

Chenyan Wu, Yandong Li, Xianfeng Tang, Zhe Huang+1

The Pennsylvania State University · Google · Amazon

3D Human Reconstruction Graph Neural Networks

Open paper AskPDF arXiv

Reconstructing multi-human body meshes from a single monocular image is a challenging problem, requiring both individual body meshes and relative 3D positions to produce a coherent scene representation. We introduce MUG (Multi-hUman Graph network), which uses a single graph neural network to construct coherent multi-human meshes from only 2D poses. Our approach demonstrates that pose-based methods can jointly perform depth reasoning and multi-human mesh generation. Existing image-based methods often rely on lab-collected datasets with accurate 3D labels, introducing domain gaps when applied to in-the-wild or art images due to differences in appearance and context. In contrast, MUG leverages the geometric consistency of 2D poses across diverse datasets to mitigate such discrepancies. MUG operates in three phases. First, it constructs a novel heterogeneous graph from multi-human 2D poses, connecting nodes across people and within individuals to capture inter-human interactions and body geometry, including skeletal and mesh structures. Second, a dual-branch graph neural network predicts inter-human depth relations and root-joint-relative mesh coordinates. Finally, these outputs are combined to construct complete multi-human 3D meshes. Despite using only 2D poses and a streamlined architecture, MUG outperforms existing multi-human mesh estimation methods across datasets from diverse domains, including RH, MuPoTS-3D, and 3DPW. Qualitative results show that MUG can handle art images, through-wall scenarios, and poor lighting conditions when paired with advanced 2D pose networks. Both qualitative and quantitative evaluations highlight its generalization ability in open-world scenarios.