Research map / 3D vision / Human pose estimation
Human Action Transfer Based on 3D Model Reconstruction
Shanyan Guan, Shuo Wen, Dexin Yang, Bingbing Ni, Wendong Zhang, Jun Tang, Xiaokang Yang
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 |
|---|---|---|---|
| Use proposed 3D-model framework instead of Variational U-net | human action transfer framework · Human3.6M | SSIM: 0.88032 -> 0.90103 | Improved |
| Use proposed 3D-model framework instead of Variational U-net | human action transfer framework · Human3.6M | PSNR: 30.26613 -> 33.98061 | Improved |
| Use proposed 3D-model framework instead of Variational U-net | human action transfer framework · HumanEva-I | SSIM: 0.87658 -> 0.89630 | Improved |
| Use proposed 3D-model framework instead of Variational U-net | human action transfer framework · HumanEva-I | PSNR: 27.21103 -> 30.69898 | Improved |
| Add temporal connection (previous pose) to pose generator (ablation) | human action transfer framework · Human3.6M | MOS: 3.89 ± 0.01 -> 4.13 ± 0.09 | Improved |
| Add temporal connection (previous pose) to pose generator (ablation) | human action transfer framework · HumanEva-I | MOS: 3.91 ± 0.06 -> 4.20 ± 0.11 | Improved |
| Add prior loss to 3D pose generator (ablation) | human action transfer framework · Human3.6M | joint reprojection accuracy (qualitative) | Improved |
| Use 3 appearance images of different views for texture extraction | human action transfer framework · Human3.6M | Brenner gradient | Improved |
| Use semantic-aware texture rendering instead of Alldieck et al. | human action transfer framework · People-Snapshot | texture quality (qualitative) | Improved |
Closest papers by meaning
- Human Motion Transfer With 3D Constraints and Detail Enhancement (TPAMI 2022 · similarity 0.91)
- Action2video: Generating Videos of Human 3D Actions (IJCV 2022 · similarity 0.88)
- HumanVid: Demystifying Training Data for Camera-controllable Human Image Animation (NeurIPS 2024 · similarity 0.85)
- Free-viewpoint Human Animation with Pose-correlated Reference Selection (CVPR 2025 · similarity 0.85)
- En3D: An Enhanced Generative Model for Sculpting 3D Humans from 2D Synthetic Data (CVPR 2024 · similarity 0.85)
- SyncHuman: Synchronizing 2D and 3D Generative Models for Single-view Human Reconstruction (NeurIPS 2025 · similarity 0.85)
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.
