Research map / 3D vision / NeRF and radiance fields
MetaCap: Meta-learning Priors from Multi-View Imagery for Sparse-view Human Performance Capture and Rendering
Guoxing Sun, Rishabh Dabral, Pascal Fua, Christian Theobalt, Marc Habermann
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 |
|---|---|---|---|
| use meta-learned weights with DDC canonicalization instead of random init (ablation) | MetaCap · DynaCap S2 novel view | PSNR: 23.566 -> 26.529 | Improved |
| use meta-learned weights instead of pre-trained weights (ablation) | MetaCap · DynaCap S2 novel view | PSNR: 25.829 -> 26.529 | Improved |
| add DDC space canonicalization to meta-learning instead of root-normalized space (ablation) | MetaCap · DynaCap S2 novel view | PSNR: 23.094 -> 26.529 | Improved |
| use DDC template instead of SMPL for canonicalization (ablation) | MetaCap · DynaCap S2 | Chamfer: 0.799 -> 0.679 | Improved |
| random init with DDC vs root space canonicalization (ablation) | MetaCap · DynaCap S2 novel view | PSNR: 23.802 -> 23.566 | No clear effect |
| add occlusion handling (OH) under monocular fine-tuning | MetaCap · monocular fine-tuning | reconstruction quality | Improved |
| use sparse 4-view mocap instead of dense 34-view mocap (ablation) | MetaCap · DynaCap S2 | PSNR: 26.529 -> 26.240 | Hurt |
| increase number of cameras | MetaCap · DynaCap S2 | rendering and reconstruction | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- ARAH — ARAH: Animatable Volume Rendering of Articulated Human SDFs (ECCV 2022) · better on 10 of 11 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Neural Human Performer: Learning Generalizable Radiance Fields for Human Performance Rendering (NeurIPS 2021 · similarity 0.89)
- SPARF: Neural Radiance Fields From Sparse and Noisy Poses (CVPR 2023 · similarity 0.88)
- H-NeRF: Neural Radiance Fields for Rendering and Temporal Reconstruction of Humans in Motion (NeurIPS 2021 · similarity 0.88)
- Sparse-view Pose Estimation and Reconstruction via Analysis by Generative Synthesis (NeurIPS 2024 · similarity 0.87)
- Sparse3D: Distilling Multiview-Consistent Diffusion for Object Reconstruction from Sparse Views (AAAI 2024 · similarity 0.87)
- MuRF: Multi-Baseline Radiance Fields (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.
