Research map / Video / Video generation
TIGER: Taming Identity, Geometry, and Generative Priors for High-Quality Face Video Restoration
Yang Zhou, Wenxue Li, Peng Zhang, Yifei Chen, Fei Wang, Daiguo Zhou
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 TIGER instead of DOVE (strongest baseline) | TIGER · VFHQ-Test | FVD: 103.892 -> 40.4222 | Improved |
| use TIGER instead of DOVE | TIGER · VFHQ-Test | LPIPS: 0.1189 -> 0.0697 | Improved |
| use TIGER instead of DOVE | TIGER · VFHQ-Test | IDS: 0.9283 -> 0.9481 | Improved |
| use TIGER instead of STAR | TIGER · VFHQ-Test | CLIP-IQA: 0.5889 -> 0.4756 | Hurt |
| add identity prior (identity embedding input) (ablation) | TIGER · VFHQ-Test | VIRD, LPIPS, FVD | Improved |
| add 3D geometry prior (normal maps) (ablation) | TIGER · VFHQ-Test | VIRD, LPIPS, FVD | Improved |
| add Stage II decoder-coupled detail refinement (ablation) | TIGER · VFHQ-Test | perceptual quality | Improved |
| add Stage III adversarial distribution alignment (ablation) | TIGER · VFHQ-Test | identity consistency, perceptual quality | Improved |
| use cross-source fusion (identity from reference, motion from degraded video) instead of single-source parameters (ablation) | TIGER · VFHQ-Test | FVR metrics | Improved |
| add identity embeddings in cross-attention versus text-only (ablation) | TIGER · VFHQ-Test | overall metrics | Improved |
| increase latent loss weight lambda1 from 1.0 to 2.0 (ablation) | TIGER · VoxCeleb2 (loss weight ablation) | VIRD: 0.9280 -> 0.9315 | Hurt |
| reduce pixel loss weight lambda3 from 10.0 to 5.0 (ablation) | TIGER · VoxCeleb2 (loss weight ablation) | VIRD: 0.9280 -> 0.9356 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 3 on most shared results.
- SVFR — SVFR: A Unified Framework for Generalized Video Face Restoration (CVPR 2025) · better on 5 of 5 shared results
- SeedVR2-3B — SeedVR: Seeding Infinity in Diffusion Transformer Towards Generic Video Restoration (CVPR 2025) · better on 0 of 1 shared result
- BasicVSR++ — BasicVSR: The Search for Essential Components in Video Super-Resolution and Beyond (CVPR 2021) · better on 1 of 1 shared result
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
- FaithfulFaces: Pose-Faithful Facial Identity Preservation for Text-to-Video Generation (NeurIPS 2026 · similarity 0.89)
- Face Video Deblurring Using 3D Facial Priors (ICCV 2019 · similarity 0.89)
- Unwarping the Lens: A Physics-Grounded Approach to Video Glasses Removal (ECCV 2026 · similarity 0.89)
- Face Restoration via Plug-and-Play 3D Facial Priors (TPAMI 2021 · similarity 0.88)
- Bridging Information Asymmetry: A Hierarchical Framework for Deterministic Blind Face Restoration (TPAMI 2026 · similarity 0.88)
- Blind Face Restoration under Extreme Conditions: Leveraging 3D-2D Prior Fusion for Superior Structural and Texture Recovery (AAAI 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.
