Research map / 3D vision / Face and shape
Deep 3D Portrait From a Single Image
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 |
|---|---|---|---|
| add face depth as a conditional input to the depth estimation network (ablation) | Ours · Biwi Kinect Head Pose Database, face | 3D reconstruction error (mm): 4.31 -> 3.88 | Helped |
| add face depth as a conditional input to the depth estimation network (ablation) | Ours · Biwi Kinect Head Pose Database, non-face | 3D reconstruction error (mm): 7.39 -> 6.78 | Helped |
| use the proposed 3D head reconstruction instead of Zhu's method | 3D head reconstruction · Biwi Kinect Head Pose Database, face | 3D reconstruction error (mm): 5.05 -> 3.88 | Helped |
| use the proposed 3D head reconstruction instead of Zhu's method | 3D head reconstruction · Biwi Kinect Head Pose Database, non-face | 3D reconstruction error (mm): 8.56 -> 6.78 | Helped |
| use the proposed pose manipulation instead of Chai's method | Portrait rotation · 10 images used in the Chai comparison | deep feature cosine similarity: 0.829 -> 0.856 | Helped |
| add layer-order loss to penalize hair depth behind occluded face (ablation) | Hair and ear depth estimation · 3D head reconstruction | reconstruction accuracy | Helped |
| regenerate the whole image instead of merging the network output with the raw image (ablation) | Image refinement network · Head pose manipulation | visual result quality | Helped |
| add adversarial loss to image refinement (ablation) | Image refinement network · Head pose manipulation | image sharpness | Helped |
| add adversarial loss to image refinement (ablation) | Image refinement network · Head pose manipulation | artifacts | Hurt |
| add perceptual feature loss for paired refinement data (ablation) | Image refinement network · Head pose manipulation | visual artifact quality | Helped |
Closest papers by meaning
- Unsupervised Depth Estimation, 3D Face Rotation and Replacement (NeurIPS 2018 · similarity 0.90)
- Generalizable One-shot 3D Neural Head Avatar (NeurIPS 2023 · similarity 0.89)
- Learning Detailed Face Reconstruction From a Single Image (CVPR 2017 · similarity 0.89)
- PanoHead: Geometry-Aware 3D Full-Head Synthesis in 360deg (CVPR 2023 · similarity 0.88)
- VOODOO 3D: Volumetric Portrait Disentanglement For One-Shot 3D Head Reenactment (CVPR 2024 · similarity 0.86)
- Do Not DeepFake Me: Privacy-Preserving Neural 3D Head Reconstruction Without Sensitive Images (AAAI 2025 · 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.