Research map / 3D vision / Human pose estimation
Pose2Mesh: Graph Convolutional Network for 3D Human Pose and Mesh Recovery from a 2D Human Pose
Hongsuk Choi, Gyeongsik Moon, Kyoung Mu Lee
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 |
|---|---|---|---|
| regress vertex coordinates with GraphCNN instead of FC network (ablation) | Pose2Mesh · Human3.6M | MPJPE: 119.6 -> 64.9 | Improved |
| regress vertex coordinates instead of SMPL parameters (GraphCNN) (ablation) | Pose2Mesh · Human3.6M | MPJPE: 79.1 -> 64.9 | Improved |
| regress vertex coordinates instead of SMPL parameters (FC network) (ablation) | FC network · Human3.6M | MPJPE: 72.8 -> 119.6 | Hurt |
| use GraphCNN instead of FC for SMPL parameter regression (ablation) | param regression · Human3.6M | MPJPE: 72.8 -> 79.1 | Hurt |
| use coarse-to-fine upsampling instead of direct upsampling (ablation) | MeshNet · Human3.6M | MPJPE: 65.3 -> 64.9 | Improved |
| use coarse-to-fine upsampling instead of direct upsampling (ablation) | MeshNet · Human3.6M | fps: 24 -> 37 fps, 10G -> 6G memory | Improved |
| concatenate 3D pose with 2D input instead of 2D alone (ablation) | Pose2Mesh · 3DPW | MPJPE: 101.1 -> 100.5 | Improved |
| feed only the PoseNet 3D pose instead of 2D pose (ablation) | Pose2Mesh · 3DPW | MPJPE: 101.1 -> 103.2 | Hurt |
| add 2D pose to 3D pose input (vs 3D pose only) (ablation) | Pose2Mesh · 3DPW | MPJPE: 103.2 -> 100.5 | Improved |
| add COCO to the Human3.6M training set | Pose2Mesh · Human3.6M test | MPJPE: 64.9 -> 67.9 | Hurt |
| add COCO to the Human3.6M training set | Pose2Mesh (HR) · 3DPW | MPJPE: 100.5 -> 91.4 | Improved |
| add COCO to training for image-based SPIN | SPIN · 3DPW | MPJPE: 313.8 -> 113.1 | Improved |
Closest papers by meaning
- Convolutional Mesh Regression for Single-Image Human Shape Reconstruction (CVPR 2019 · similarity 0.92)
- Learning to Estimate 3D Human Pose and Shape From a Single Color Image (CVPR 2018 · similarity 0.91)
- A Simple yet Effective Baseline for 3D Human Pose Estimation (ICCV 2017 · similarity 0.91)
- Locally Connected Network for Monocular 3D Human Pose Estimation (TPAMI 2020 · similarity 0.90)
- Optimizing Network Structure for 3D Human Pose Estimation (ICCV 2019 · similarity 0.90)
- 3D Human Pose Estimation From a Single Image via Distance Matrix Regression (CVPR 2017 · similarity 0.90)
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.
