Research map / Video / Video generation
CelebV-HQ: A Large-Scale Video Facial Attributes Dataset
Hao Zhu, Wayne Wu, Wentao Zhu, Liming Jiang, Siwei Tang, Li Zhang, Ziwei Liu, Chen Change Loy
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 temporal optical-flow L2 constraint | StarGAN-v2 (Reference) · CelebV-HQ editing, Gender | FVD: 284.80 -> 262.01 | Improved |
| add temporal optical-flow L2 constraint | StarGAN-v2 (Label) · CelebV-HQ editing, Gender | FVD: 258.36 -> 189.40 | Improved |
| add temporal optical-flow L2 constraint | MUNIT · CelebV-HQ editing, Gender | FVD: 219.96 -> 211.45 | Improved |
| add temporal optical-flow L2 constraint | MUNIT · CelebV-HQ editing, Eyeglasses | FVD: 204.12 -> 158.87 | Improved |
| add temporal optical-flow L2 constraint | StarGAN-v2 (Reference) · CelebV-HQ editing, Gender | FID: 80.61 -> 82.99 | Hurt |
| add temporal optical-flow L2 constraint | StarGAN-v2 (Label) · CelebV-HQ editing, Gender | FID: 65.70 -> 55.73 | Improved |
| add temporal optical-flow L2 constraint | StarGAN-v2 (Brown Hair, Reference) · CelebV-HQ editing | FID: 77.26 -> 89.07 | Hurt |
| add temporal optical-flow L2 constraint | MUNIT · CelebV-HQ editing, Eyeglasses | FID: 30.65 -> 31.23 | No clear effect |
| train on CelebV-HQ instead of automatically labeled Vox | StarGAN-v2 (Gender, Label) · video facial editing | FVD: 629.09 -> 189.04 | Improved |
| train on CelebV-HQ instead of automatically labeled Vox | StarGAN-v2 (Brown hair, Label) · video facial editing | FVD: 500.77 -> 232.63 | Improved |
| use StyleGAN-V instead of VideoGPT | unconditional generation · CelebV-HQ | FVD: 177.89 -> 69.17 | Improved |
| use MoCoGAN-HD instead of StyleGAN-V | unconditional generation · FaceForensics | FID: 9.45 -> 7.12 | Improved |
Closest papers by meaning
- CelebV-Text: A Large-Scale Facial Text-Video Dataset (CVPR 2023 · similarity 0.92)
- DH-FaceVid-1K: A Large-Scale High-Quality Dataset for Face Video Generation (ICCV 2025 · similarity 0.88)
- MS-Celeb-1M: A Dataset and Benchmark for Large-Scale Face Recognition (ECCV 2016 · similarity 0.88)
- Synergizing Motion and Appearance: Multi-Scale Compensatory Codebooks for Talking Head Video Generation (CVPR 2025 · similarity 0.83)
- OpenS2V-Nexus: A Detailed Benchmark and Million-Scale Dataset for Subject-to-Video Generation (NeurIPS 2025 · similarity 0.83)
- MVHumanNet: A Large-scale Dataset of Multi-view Daily Dressing Human Captures (CVPR 2024 · similarity 0.82)
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.
