Research map / 3D vision / Face and shape
Textured Neural Avatars
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Represent appearance with explicit learned textures and predict texture coordinates instead of directly translating pose to RGB | Textured Neural Avatars · CMU Panoptic hold-out cameras and motions | SSIM: 0.908 -> 0.919; 0.916 -> 0.922; 0.905 -> 0.914; 0.918 -> 0.920 | Helped |
| Represent appearance with explicit learned textures and predict texture coordinates instead of directly translating pose to RGB | Textured Neural Avatars · CMU Panoptic hold-out cameras and motions | Frechet distance: 6.7 -> 8.8; 7.0 -> 10.7; 7.7 -> 8.9; 7.0 -> 10.4 | Hurt |
| Replace direct RGB-and-mask prediction with textured body-part assignments and coordinates (ablation) | Direct baseline · CMU Panoptic hold-out cameras and motions | SSIM: 0.899 -> 0.919; 0.907 -> 0.922; 0.896 -> 0.914; 0.907 -> 0.920 | Helped |
| Replace direct RGB-and-mask prediction with textured body-part assignments and coordinates (ablation) | Direct baseline · CMU Panoptic hold-out cameras and motions | Frechet distance: 7.3 -> 8.8; 8.8 -> 10.7; 10.7 -> 8.9; 9.7 -> 10.4 | Mixed |
| Add DensePose outputs as input alongside keypoints (ablation) | V2V on CMU1-6 · CMU1-6 | considered metrics | No clear effect |
| Use the textured avatar rather than V2V when fine-tuning from a single video | Single-video avatar · Hold-out camera and motion, two subjects | user preference rate: 55%; 65% | Helped |
Closest papers by meaning
- Pose Modulated Avatars from Video (ICLR 2024 · similarity 0.89)
- VINECS: Video-based Neural Character Skinning (CVPR 2024 · similarity 0.88)
- SCULPT: Shape-Conditioned Unpaired Learning of Pose-dependent Clothed and Textured Human Meshes (CVPR 2024 · similarity 0.88)
- RAM-Avatar: Real-time Photo-Realistic Avatar from Monocular Videos with Full-body Control (CVPR 2024 · similarity 0.88)
- NECA: Neural Customizable Human Avatar (CVPR 2024 · similarity 0.87)
- TexVocab: Texture Vocabulary-conditioned Human Avatars (CVPR 2024 · similarity 0.87)
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.