Research map / 3D vision / NeRF and radiance fields
Real-Time Neural Character Rendering with Pose-Guided Multiplane Images
Hao Ouyang, Bo Zhang, Pan Zhang, Hao Yang, Jiaolong Yang, Dong Chen, Qifeng Chen, Fang Wen
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 adaptive (learnable) MPI plane depth (ablation) | Pose-guided MPI · Captured character sequences 3-5 ablation (mean) | PSNR: 20.54 -> 25.17 | Improved |
| add adaptive (learnable) MPI plane depth (ablation) | Pose-guided MPI · Captured character sequences 3-5 ablation (mean) | LPIPS: 0.210 -> 0.118 | Improved |
| add learnable exposure coefficients (ablation) | Pose-guided MPI · Captured character sequences 3-5 ablation (mean) | PSNR: 25.02 -> 25.17 | Improved |
| add learnable camera poses (ablation) | Pose-guided MPI · Captured character sequences 3-5 ablation (mean) | PSNR: 25.17 -> 25.54 | Improved |
| add learnable camera poses (ablation) | Pose-guided MPI · Captured character sequence 5 (5000 images) ablation | PSNR: 21.92 -> 23.12 | Improved |
| use full-size U-Net instead of smaller U-Net (Ours-s) (ablation) | Pose-guided MPI · Captured character sequences 3-5 ablation (mean) | PSNR: 24.77 -> 25.17 | Improved |
| use full-size U-Net instead of smaller U-Net (Ours-s) (ablation) | Pose-guided MPI · Inference speed, 360x640, single V100 | FPS: 32.26 -> 20.41 | Hurt |
| pose-guided MPI versus HyperNeRF on PSNR | HyperNeRF · Captured character sequences 1-3 (mean) | PSNR: 27.48 -> 27.29 | Hurt |
| pose-guided MPI versus HyperNeRF on LPIPS | HyperNeRF · Captured character sequences 1-3 (mean) | LPIPS: 0.178 -> 0.079 | Improved |
| pose-guided MPI versus HyperNeRF on SSIM | HyperNeRF · Captured character sequences 1-3 (mean) | SSIM: 0.942 -> 0.952 | Improved |
| pose-guided MPI versus static MPI NeX | NeX · Captured character sequences 1-3 (mean) | PSNR: 23.65 -> 27.29 | Improved |
| pose-guided MPI versus HyperNeRF on speed | HyperNeRF · Inference speed, 360x640, single V100 | FPS: 0.051 -> 20.41 | Improved |
Closest papers by meaning
- Pushing the Boundaries of View Extrapolation With Multiplane Images (CVPR 2019 · similarity 0.86)
- Holoported Characters: Real-time Free-viewpoint Rendering of Humans from Sparse RGB Cameras (CVPR 2024 · similarity 0.85)
- MIMO: Controllable Character Video Synthesis with Spatial Decomposed Modeling (CVPR 2025 · similarity 0.84)
- Free-viewpoint Human Animation with Pose-correlated Reference Selection (CVPR 2025 · similarity 0.84)
- AniGS: Animatable Gaussian Avatar from a Single Image with Inconsistent Gaussian Reconstruction (CVPR 2025 · similarity 0.83)
- CAP4D: Creating Animatable 4D Portrait Avatars with Morphable Multi-View Diffusion Models (CVPR 2025 · similarity 0.83)
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.
