Research map / 3D vision / Human pose estimation
3D Human Texture Estimation From a Single Image With Transformers
Xiangyu Xu, Chen Change Loy
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 Transformer units (ablation) | Texformer · Market-1501 test (100 identities) | LPIPS: 0.1235 -> 0.1154 | Improved |
| add Transformer units (ablation) | Texformer · Market-1501 test (100 identities) | CosSim: 0.5413 -> 0.5747 | Improved |
| add Transformer units (ablation) | Texformer · Market-1501 test (100 identities) | SSIM: 0.7242 -> 0.7422 | Improved |
| add multi-scale feature fusion (ablation) | Texformer · Market-1501 test (100 identities) | LPIPS: 0.1181 -> 0.1154 | Improved |
| add multi-scale feature fusion (ablation) | Texformer · Market-1501 test (100 identities) | CosSim: 0.5715 -> 0.5747 | Improved |
| add mask-fusion of RGB and texture flow to RGB-only output (ablation) | Texformer · Market-1501 test (100 identities) | SSIM: 0.7167 -> 0.7422 | Improved |
| add mask-fusion of RGB and texture flow to RGB-only output (ablation) | Texformer · Market-1501 test (100 identities) | LPIPS: 0.1261 -> 0.1154 | Improved |
| add mask-fusion of RGB and texture flow to flow-only output (ablation) | Texformer · Market-1501 test (100 identities) | CosSim: 0.5515 -> 0.5747 | Improved |
| add mask-fusion of RGB and texture flow to flow-only output (ablation) | Texformer · Market-1501 test (100 identities) | LPIPS: 0.1027 -> 0.1154 | Hurt |
| add part-style loss (ablation) | Texformer · Market-1501 test (100 identities) | LPIPS: 0.1412 -> 0.1154 | Improved |
| add part-style loss (ablation) | Texformer · Market-1501 test (100 identities) | SSIM: 0.7158 -> 0.7422 | Improved |
| add part-style loss (ablation) | Texformer · Market-1501 test (100 identities) | CosSim: 0.5783 -> 0.5747 | Hurt |
Closest papers by meaning
- Refining 3D Human Texture Estimation From a Single Image (TPAMI 2024 · similarity 0.90)
- Learning to Estimate 3D Human Pose and Shape From a Single Color Image (CVPR 2018 · similarity 0.85)
- Moulding Humans: Non-Parametric 3D Human Shape Estimation From Single Images (ICCV 2019 · similarity 0.85)
- Image-Based 3D Object Reconstruction: State-of-the-Art and Trends in the Deep Learning Era (TPAMI 2019 · similarity 0.84)
- Towards High-Fidelity 3D Face Reconstruction From In-the-Wild Images Using Graph Convolutional Networks (CVPR 2020 · similarity 0.84)
- Texture Fields: Learning Texture Representations in Function Space (ICCV 2019 · similarity 0.84)
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.
