Research map / 3D vision / Human pose estimation
SemGAN: A Semantic and Hierarchical Adversarial Network for 3D Human Pose Estimation
Haodong Feng, Yu Xin, Guoqing Li
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 the two-stage GAN with ECF-Discriminator and SPH-Generator over explicit-loss-only baseline (ablation) | SPH-Generator · Human3.6M (CPN 2D input, T=243) | MPJPE: 51.2 -> 39.7 | Improved |
| add the SPH-Generator (ablation) | SemGAN · Human3.6M (CPN 2D input, T=243) | MPJPE: 42.4 -> 39.7 (table); text says 45.4 -> 43.9 | Mixed |
| add the frequency branch to discriminator (ablation) | SemGAN · Human3.6M (CPN 2D input, T=243) | MPJPE: 44.2 -> 39.7 (table); text says +0.3mm | Mixed |
| add the time branch to discriminator (ablation) | SemGAN · Human3.6M (CPN 2D input, T=243) | MPJPE: 44.5 -> 39.7 (table); text says +0.6mm | Mixed |
| use non-uniform 10/20/70 frequency bands instead of uniform split (ablation) | ECF-Discriminator · Human3.6M (CPN 2D input, T=243) | MPJPE: 40.5 -> 39.7 | Improved |
| use 10/20/70 split instead of high-frequency bias 5/10/85 (ablation) | ECF-Discriminator · Human3.6M (CPN 2D input, T=243) | MPJPE: 41.5 -> 39.7 | Improved |
| use hierarchical semantic generator and cross-frequency constraints versus parameter-matched flat Transformer (ablation) | SemGAN · Human3.6M (CPN) | MPJPE: 40.4 -> 39.7 | Improved |
| SemGAN versus KTPFormer at T=243 on GT 2D input | SemGAN · Human3.6M (GT 2D input, T=243) | MPJPE: 19.0 -> 18.8 | Improved |
| SemGAN versus KTPFormer on MPI-INF-3DHP T=81 | SemGAN · MPI-INF-3DHP (GT 2D, T=81) | PCK: 98.9 -> 98.7 | Hurt |
| SemGAN versus KTPFormer on MPI-INF-3DHP T=81 | SemGAN · MPI-INF-3DHP (GT 2D, T=81) | MPJPE: 16.7 -> 16.5 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- KTPFormer — KTPFormer: Kinematics and Trajectory Prior Knowledge-Enhanced Transformer for 3D Human Pose Estimation (CVPR 2024) · better on 7 of 8 shared results
- MixSTE — MixSTE: Seq2seq Mixed Spatio-Temporal Encoder for 3D Human Pose Estimation in Video (CVPR 2022) · better on 4 of 4 shared results
- GLA-GCN — GLA-GCN: Global-local Adaptive Graph Convolutional Network for 3D Human Pose Estimation from Monocular Video (ICCV 2023) · 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
- 3D Human Pose Estimation in the Wild by Adversarial Learning (CVPR 2018 · similarity 0.90)
- A Simple yet Effective Baseline for 3D Human Pose Estimation (ICCV 2017 · similarity 0.90)
- 3D Human Pose Estimation From a Single Image via Distance Matrix Regression (CVPR 2017 · similarity 0.90)
- Adversarial PoseNet: A Structure-Aware Convolutional Network for Human Pose Estimation (ICCV 2017 · similarity 0.89)
- Semantic Graph Convolutional Networks for 3D Human Pose Regression (CVPR 2019 · similarity 0.89)
- A Dual-Augmentor Framework for Domain Generalization in 3D Human Pose Estimation (CVPR 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.
