Research map / 3D vision / Face and shape
ANR: Articulated Neural Rendering for Virtual 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 |
|---|---|---|---|
| replace DNR with ANR's articulated rendering framework | DNR avatar rendering · Novel pose synthesis | SSIM: 0.9398 -> 0.9738 | Helped |
| replace DNR with ANR's articulated rendering framework | DNR avatar rendering · Novel pose synthesis | LPIPS: 0.0918 -> 0.0508 | Helped |
| replace DNR with ANR's articulated rendering framework | DNR avatar rendering · Novel pose synthesis | mFID: 92 -> 74 | Helped |
| add feature loss to pixel loss (ablation) | ANR · ANR loss ablation | LPIPS: 0.086 -> 0.065 | Helped |
| add feature loss to pixel loss (ablation) | ANR · ANR loss ablation | FLIP: 0.029 -> 0.033 | Hurt |
| add TV loss to pixel and feature losses (ablation) | ANR · ANR loss ablation | LPIPS: 0.065 -> 0.064 | Helped |
| use two rendering stages instead of one (ablation) | ANR · ANR model ablation | LPIPS: 0.063 -> 0.050 | Helped |
| add split optimization using keyframes for texture learning (ablation) | two-stage ANR · ANR model ablation | LPIPS: 0.058 -> 0.050 | Helped |
| add normal information between the rendering stages (ablation) | ANR · Avatar rendering | level of detail | Helped |
| increase U-Net capacity without changing the rendering design (ablation) | DNR-style neural renderer · Avatar rendering | generation quality | No clear effect |
| use pixel loss alone instead of mask and feature losses (ablation) | ANR · Avatar rendering | avatar shape plausibility | Hurt |
| train with keyframes for texture and remaining frames for rendering (ablation) | ANR · Avatar training | convergence speed: 5X faster | Helped |
Closest papers by meaning
- 3DGS-Avatar: Animatable Avatars via Deformable 3D Gaussian Splatting (CVPR 2024 · similarity 0.89)
- Artist-Friendly Relightable and Animatable Neural Heads (CVPR 2024 · similarity 0.88)
- Relightable and Animatable Neural Avatars from Videos (AAAI 2024 · similarity 0.87)
- Efficient 3D Implicit Head Avatar with Mesh-anchored Hash Table Blendshapes (CVPR 2024 · similarity 0.87)
- CAP4D: Creating Animatable 4D Portrait Avatars with Morphable Multi-View Diffusion Models (CVPR 2025 · similarity 0.87)
- 3DToonify: Creating Your High-Fidelity 3D Stylized Avatar Easily from 2D Portrait Images (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.