Research map / 3D vision / NeRF and radiance fields
Structured Local Radiance Fields for Human Avatar Modeling
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 |
|---|---|---|---|
| Use structured local radiance fields with node residual translations and dynamic detail embeddings instead of Animatable NeRF | Clothed human avatar modeling · Hoody, Jacket, Dress1, and Dress2 novel-pose synthesis | PSNR: 22.43 -> 24.94; 24.30 -> 25.24; 19.52 -> 23.43; 20.49 -> 22.19 | Helped |
| Use structured local radiance fields with node residual translations and dynamic detail embeddings instead of Animatable NeRF | Clothed human avatar modeling · Hoody, Jacket, Dress1, and Dress2 novel-pose synthesis | SSIM: 0.893 -> 0.928; 0.909 -> 0.927; 0.848 -> 0.891; 0.877 -> 0.900 | Helped |
| Use structured local radiance fields with pose-conditioned node variables instead of Neural Body | Clothed human avatar modeling · ZJU-MoCap 387 and 392, seen and unseen poses | PSNR: 25.79 -> 28.32; 21.60 -> 23.61; 29.44 -> 30.79; 25.76 -> 26.74 | Helped |
| Use structured local radiance fields with pose-conditioned node variables instead of Neural Body | Clothed human avatar modeling · ZJU-MoCap 387 and 392, seen and unseen poses | SSIM: 0.928 -> 0.953; 0.870 -> 0.905; 0.946 -> 0.958; 0.909 -> 0.927 | Helped |
| Add node residual translations and dynamic detail embeddings (ablation) | Structured local radiance fields · Dress sequence, training poses | PSNR: 17.52 -> 21.48 | Helped |
| Add dynamic detail embeddings (ablation) | Structured local radiance fields with node residual translations · Dress sequence, training poses | PSNR: 20.47 -> 21.48 | Helped |
| Replace cVAE with deterministic regression from pose (ablation) | Node-related variable prediction · Training frames | PSNR: 25.62 -> 25.34 | Hurt |
| Replace the time-stamp input with learnable per-frame latent codes (ablation) | cVAE node-related variable prediction · Node motion trajectories | temporal smoothness | Hurt |
Closest papers by meaning
- Pose Modulated Avatars from Video (ICLR 2024 · similarity 0.90)
- Animatable Gaussians: Learning Pose-dependent Gaussian Maps for High-fidelity Human Avatar Modeling (CVPR 2024 · similarity 0.90)
- 3DGS-Avatar: Animatable Avatars via Deformable 3D Gaussian Splatting (CVPR 2024 · similarity 0.90)
- Human Gaussian Splatting: Real-time Rendering of Animatable Avatars (CVPR 2024 · similarity 0.90)
- Relightable and Animatable Neural Avatars from Videos (AAAI 2024 · similarity 0.89)
- Relightable and Animatable Neural Avatar from Sparse-View Video (CVPR 2024 · similarity 0.88)
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.