Research map / 3D vision / Human pose estimation
Intrinsic-Extrinsic Preserved GANs for Unsupervised 3D Pose Transfer
Haoyu Chen, Hao Tang, Henglin Shi, Wei Peng, Nicu Sebe, Guoying Zhao
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 Laplacian co-occurrence discriminator (3 reference Laplacians) (ablation) | IEP-GAN · FAUST | disentanglement error: 2.6 -> 0.19 | Improved |
| increase reference Laplacians from 1 to 3 (ablation) | IEP-GAN · FAUST | disentanglement error: 0.59 -> 0.19 | Improved |
| increase reference Laplacians from 2 to 3 (ablation) | IEP-GAN · FAUST | disentanglement error: 0.21 -> 0.19 | No clear effect |
| add regional geodesic preservation loss (ablation) | IEP-GAN · FAUST | interpolation error: 3.11e-2 -> 4.02e-3 | Improved |
| use adaptive instead of random sub-region sampling for regional geodesic loss (ablation) | IEP-GAN · FAUST | interpolation error: 7.29e-3 -> 4.02e-3 | Improved |
| use regional adaptive geodesic loss instead of global geodesic loss (ablation) | IEP-GAN · FAUST | runtime (s): 8.31 s -> 2.07 s | Improved |
| use regional adaptive geodesic loss instead of global geodesic loss (ablation) | IEP-GAN · FAUST | interpolation error: 3.94e-3 -> 4.02e-3 | No clear effect |
| IEP-GAN unsupervised vs supervised LIMP Geo on interpolation error | IEP-GAN · FAUST | interpolation error: 3.82e-3 -> 4.02e-3 | Hurt |
| IEP-GAN vs LIMP Geo on disentanglement error | IEP-GAN · FAUST | disentanglement error: 3.48 -> 0.19 | Improved |
| IEP-GAN vs LIMP Geo on disentanglement error | IEP-GAN · DFAUST | disentanglement error: 4.11 -> 0.34 | Improved |
Closest papers by meaning
- Towards Robust 3D Pose Transfer with Adversarial Learning (CVPR 2024 · similarity 0.90)
- Unsupervised 3D Pose Transfer With Cross Consistency and Dual Reconstruction (TPAMI 2023 · similarity 0.89)
- Unsupervised Learning of Efficient Geometry-Aware Neural Articulated Representations (ECCV 2022 · similarity 0.88)
- HoloGAN: Unsupervised Learning of 3D Representations From Natural Images (ICCV 2019 · similarity 0.87)
- Do 2D GANs Know 3D Shape? Unsupervised 3D Shape Reconstruction from 2D Image GANs (ICLR 2021 · similarity 0.87)
- Dual Encoder GAN Inversion for High-Fidelity 3D Head Reconstruction from Single Images (NeurIPS 2024 · similarity 0.86)
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.
