Research map / 3D vision / Face and shape
Face Video Deblurring Using 3D Facial Priors
Wenqi Ren, Jiaolong Yang, Senyou Deng, David Wipf, Xiaochun Cao, Xin Tong
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 the rendered 3D face as guidance (no identity vector) (ablation) | Face video deblurring network · 300VW (9 synthetic test videos) | PSNR: 36.74 -> 37.53 | Improved |
| add the rendered 3D face as guidance (no identity vector) (ablation) | Face video deblurring network · 300VW (9 synthetic test videos) | SSIM: 0.9817 -> 0.9869 | Improved |
| add the identity vector alone (no rendered face) (ablation) | Face video deblurring network · 300VW (9 synthetic test videos) | PSNR: 36.74 -> 36.82 | No clear effect |
| add the identity vector alone (no rendered face) (ablation) | Face video deblurring network · 300VW (9 synthetic test videos) | identity similarity: 0.8352 -> 0.8335 | Hurt |
| add the identity vector on top of the rendered face (ablation) | Face video deblurring network · 300VW (9 synthetic test videos) | PSNR: 37.53 -> 37.70 | Improved |
| add the identity vector on top of the rendered face (ablation) | Face video deblurring network · 300VW (9 synthetic test videos) | SSIM: 0.9869 -> 0.9849 | Hurt |
| add the identity vector on top of the rendered face (ablation) | Face video deblurring network · 300VW (9 synthetic test videos) | identity similarity: 0.8364 -> 0.8373 | Improved |
| use the proposed 3D-prior network instead of Su et al. video deblurring baseline | Face video deblurring network · 300VW (9 synthetic test videos) | PSNR: 34.40 -> 37.70 | Improved |
| use the proposed 3D-prior network instead of Su et al. on unseen data | Face video deblurring network · VidTIMIT (11 synthetic test videos) | PSNR: 37.00 -> 38.16 | Improved |
| compare running time with Su et al. at the larger resolution | Face video deblurring network · Running time 580x610 | seconds per frame: 0.77s -> 0.70s | Improved |
Closest papers by meaning
- Live Face De-Identification in Video (ICCV 2019 · similarity 0.90)
- Face Restoration via Plug-and-Play 3D Facial Priors (TPAMI 2021 · similarity 0.90)
- TIGER: Taming Identity, Geometry, and Generative Priors for High-Quality Face Video Restoration (ECCV 2026 · similarity 0.89)
- Exploiting Semantics for Face Image Deblurring (IJCV 2020 · similarity 0.89)
- Blind Face Restoration under Extreme Conditions: Leveraging 3D-2D Prior Fusion for Superior Structural and Texture Recovery (AAAI 2024 · similarity 0.86)
- Efficient Video Face Enhancement with Enhanced Spatial-Temporal Consistency (CVPR 2025 · similarity 0.85)
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.
