Research map / 3D vision / NeRF and radiance fields
Neural Human Performer: Learning Generalizable Radiance Fields for Human Performance Rendering
Read the paperCheck 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 the temporal and multi-view transformer architecture instead of generalizable NeRF baselines | Neural Human Performer · ZJU-MoCap, unseen identities and poses | PSNR: 23.17 -> 24.75 | Helped |
| Use the temporal and multi-view transformer architecture instead of generalizable NeRF baselines | Neural Human Performer · AIST, unseen identities and poses | PSNR: 18.06 -> 19.03 | Helped |
| Use Neural Human Performer rather than Neural Body | ZJU-MoCap, source identities and unseen poses · ZJU-MoCap, source identities, unseen poses | PSNR: 23.79 -> 26.94 | Helped |
| Use a generalizable model trained on source identities rather than per-person optimization | ZJU-MoCap, target identities and unseen poses · ZJU-MoCap, target identities, unseen poses | PSNR: 22.88 -> 24.75 | Helped |
| Add skeletal features to pixel-aligned features (ablation) | Neural Human Performer · ZJU-MoCap, unseen identities and poses | PSNR: 22.58 -> 23.47 | Helped |
| Add pixel-aligned features to skeletal features (ablation) | Neural Human Performer · ZJU-MoCap, unseen identities and poses | PSNR: 22.31 -> 23.47 | Helped |
| Add the temporal transformer to skeletal and pixel-aligned features (ablation) | Neural Human Performer · ZJU-MoCap, unseen identities and poses | PSNR: 23.47 -> 24.21 | Helped |
| Add the multi-view transformer to skeletal and pixel-aligned features (ablation) | Neural Human Performer · ZJU-MoCap, unseen identities and poses | PSNR: 23.47 -> 24.44 | Helped |
| Add the multi-view transformer to a model with the temporal transformer (ablation) | Neural Human Performer · ZJU-MoCap, unseen identities and poses | PSNR: 24.21 -> 24.75 | Helped |
| Increase the number of input views from one to two (ablation) | Neural Human Performer · ZJU-MoCap, one timestep | PSNR: 20.13 -> 21.82 | Helped |
| Increase the number of input views from two to three (ablation) | Neural Human Performer · ZJU-MoCap, one timestep | PSNR: 21.82 -> 23.33 | Helped |
| Increase the number of input views from three to four (ablation) | Neural Human Performer · ZJU-MoCap, one timestep | PSNR: 23.33 -> 23.51 | Helped |
Closest papers by meaning
- H-NeRF: Neural Radiance Fields for Rendering and Temporal Reconstruction of Humans in Motion (NeurIPS 2021 · similarity 0.92)
- A-NeRF: Articulated Neural Radiance Fields for Learning Human Shape, Appearance, and Pose (NeurIPS 2021 · similarity 0.91)
- GM-NeRF: Learning Generalizable Model-Based Neural Radiance Fields From Multi-View Images (CVPR 2023 · similarity 0.91)
- Joint Optimization of Neural Radiance Fields and Continuous Camera Motion from a Monocular Video (CVPR 2025 · similarity 0.87)
- PoseGen: Learning to Generate 3D Human Pose Dataset with NeRF (AAAI 2024 · similarity 0.86)
- Learning Robust Generalizable Radiance Field with Visibility and Feature Augmented Point Representation (ICLR 2024 · similarity 0.86)
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.