Research map / 3D vision / Face and shape
DeepWrinkles: Accurate and Realistic Clothing Modeling
Zorah Lahner, Daniel Cremers, Tony Tung
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 sum-based temporal consistency term instead of none (ablation) | cGAN normal map generator · T-shirt 4D scan normal maps (test) | pixel-wise data error: 5.06 -> 3.72 | Improved |
| add the sum-based temporal consistency term instead of none (ablation) | cGAN normal map generator · T-shirt 4D scan normal maps (test) | pixel-wise temporal error: 5.52 -> 4.2 | Improved |
| use sum-based temporal term instead of L1 temporal term (ablation) | cGAN normal map generator · T-shirt 4D scan normal maps (test) | pixel-wise data error: 4.63 -> 3.72 | Improved |
| use sum-based temporal term instead of L2 temporal term (ablation) | cGAN normal map generator · T-shirt 4D scan normal maps (test) | pixel-wise data error: 5.16 -> 3.72 | Improved |
| use registration instead of 500-function blend shape as input normal map (ablation) | cGAN normal map generator · T-shirt 4D scan normal maps (test) | pixel-wise data error: 6.86 -> 3.72 | Improved |
| use 500 instead of 200 blend shape basis functions for input (ablation) | cGAN normal map generator · T-shirt 4D scan normal maps (test) | pixel-wise data error: 6.40 -> 6.86 | Hurt |
| use regression reconstruction instead of 500 blend shapes as input (ablation) | cGAN normal map generator · T-shirt 4D scan normal maps (test) | pixel-wise data error: 6.86 -> 7.6 | Hurt |
| use LSTM encoder-decoder instead of linear regression for pose-to-shape | pose-to-shape prediction · Pose-to-shape prediction (9213 frames) | MSE: 2.902 -> 1.892 | Improved |
| increase blend shapes from 500 to 1000 in linear regression (ablation) | pose-to-shape regression · Pose-to-shape prediction (9213 frames) | MSE: 2.902 -> 3.114 | Hurt |
Closest papers by meaning
- High-Fidelity Modeling of Generalizable Wrinkle Deformation (ECCV 2024 · similarity 0.94)
- Garment Recovery with Shape and Deformation Priors (CVPR 2024 · similarity 0.91)
- TailorNet: Predicting Clothing in 3D as a Function of Human Pose, Shape and Garment Style (CVPR 2020 · similarity 0.91)
- SCALE: Modeling Clothed Humans with a Surface Codec of Articulated Local Elements (CVPR 2021 · similarity 0.91)
- FreeCloth: Free-form Generation Enhances Challenging Clothed Human Modeling (CVPR 2025 · similarity 0.90)
- Design2Cloth: 3D Cloth Generation from 2D Masks (CVPR 2024 · similarity 0.89)
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.
