Research map / 3D vision / Human pose estimation
GLA-GCN: Global-local Adaptive Graph Convolutional Network for 3D Human Pose Estimation from Monocular Video
Bruce X.B. Yu, Zhi Zhang, Yongxu Liu, Sheng-hua Zhong, Yan Liu, Chang Wen Chen
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 AGCN blocks instead of ST-GCN blocks (ablation) | GLA-GCN · Human3.6M (GT 2D, T=27) | P-MPJPE (mm): 28.0 -> 25.8 | Improved |
| use strided temporal convolution instead of average pooling over time (ablation) | GLA-GCN · Human3.6M (GT 2D, T=27) | P-MPJPE (mm): 30.6 -> 25.8 | Improved |
| use individual connected layer instead of fully connected layer (ablation) | GLA-GCN · HumanEva-I (GT 2D, T=27) | P-MPJPE (mm): 12.4 -> 9.2 | Improved |
| break pose structure by swapping left and right limbs of the input (ablation) | GLA-GCN · HumanEva-I (GT 2D, T=27) | P-MPJPE (mm): 9.2 -> 12.4 | Hurt |
| add intermediate 3D pose sequence reconstruction loss (ablation) | GLA-GCN · Human3.6M Protocol#1 (GT 2D, T=243) | MPJPE (mm): 28.5 -> 21.0 | Improved |
| increase temporal receptive field from T=27 to T=243 (ablation) | GLA-GCN · Human3.6M Protocol#1 (GT 2D) | MPJPE (mm): 34.4 -> 28.5 | Improved |
| use GT 2D pose rather than estimated CPN 2D pose input | GLA-GCN · Human3.6M Protocol#1 (T=243) | MPJPE (mm): 44.4 -> 28.5 | Improved |
| increase output channels from 96 to 512 with intermediate 3D reconstruction (ablation) | GLA-GCN · Human3.6M Protocol#1 (GT 2D, T=243) | MPJPE (mm): 24.5 -> 21.0 | Improved |
Later papers that compared with this method
3 later accepted papers report GLA-GCN as a baseline in their results tables; 3 of them report a better number on most of the results they share with it.
- StructMamPose: From Sequential Perception to Structural Reasoning for 3D Human Pose Estimation (ICML 2026 · better on 1 of 1 shared result)
- SemGAN: A Semantic and Hierarchical Adversarial Network for 3D Human Pose Estimation (ECCV 2026 · better on 1 of 1 shared result)
- FinePOSE: Fine-Grained Prompt-Driven 3D Human Pose Estimation via Diffusion Models (CVPR 2024 · better on 3 of 3 shared results)
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
- Locally Connected Network for Monocular 3D Human Pose Estimation (TPAMI 2020 · similarity 0.93)
- Optimizing Network Structure for 3D Human Pose Estimation (ICCV 2019 · similarity 0.91)
- Modulated Graph Convolutional Network for 3D Human Pose Estimation (ICCV 2021 · similarity 0.90)
- Graph and Temporal Convolutional Networks for 3D Multi-person Pose Estimation in Monocular Videos (AAAI 2021 · similarity 0.90)
- GCNext: Towards the Unity of Graph Convolutions for Human Motion Prediction (AAAI 2024 · similarity 0.87)
- PoseMamba: Monocular 3D Human Pose Estimation with Bidirectional Global-Local Spatio-Temporal State Space Model (AAAI 2025 · similarity 0.87)
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.
