Research map / Image generation / Editing and style transfer
HyperReenact: One-Shot Reenactment via Jointly Learning to Refine and Retarget Faces
Stella Bounareli, Christos Tzelepis, Vasileios Argyriou, Ioannis Patras, Georgios Tzimiropoulos
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 |
|---|---|---|---|
| train with curriculum learning (inversion, then self, then cross-subject reenactment) (ablation) | HyperReenact · VoxCeleb1 self reenactment | AED: 6.7 -> 5.1 | Improved |
| add curriculum learning (ablation) | HyperReenact · VoxCeleb1 cross-subject reenactment | CSIM: 0.72 -> 0.68 | Hurt |
| add cross-subject reenactment training (ablation) | HyperReenact · VoxCeleb1 cross-subject reenactment | CSIM: 0.53 -> 0.68 | Improved |
| add gaze loss (ablation) | HyperReenact · VoxCeleb1 real image inversion | Gaze Error: 0.34 -> 0.24 | Improved |
| add gaze loss (ablation) | HyperReenact · VoxCeleb1 cross-subject reenactment | Gaze Error: 0.40 -> 0.31 | Improved |
| hypernetwork refinement plus pose retargeting vs prior methods on extreme pose pairs | HyperReenact · VoxCeleb1 large head pose differences | CSIM: 0.53 -> 0.58 vs Rome | Improved |
| use shared reenactment blocks for first seven layers | hypernetwork · model size | trainable parameters: 1.2B -> 300M | Improved |
| HyperReenact vs Rome on expression distance in cross-subject reenactment | HyperReenact · VoxCeleb1 cross-subject reenactment | AED: 9.3 vs 8.8 (Rome) | No clear effect |
Later papers that compared with this method
1 later accepted paper reports HyperReenact as a baseline in their results tables; 0 of them report a better number on most of the results they share with it.
- Face Adapter for Pre-Trained Diffusion Models with Fine-Grained ID and Attribute Control (ECCV 2024 · better on 1 of 2 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
- Realistic Face Reenactment via Self-Supervised Disentangling of Identity and Pose (AAAI 2020 · similarity 0.91)
- One-Shot Neural Face Reenactment via Finding Directions in GAN’s Latent Space (IJCV 2024 · similarity 0.91)
- Synthetic Prior for Few-Shot Drivable Head Avatar Inversion (CVPR 2025 · similarity 0.87)
- VOODOO 3D: Volumetric Portrait Disentanglement For One-Shot 3D Head Reenactment (CVPR 2024 · similarity 0.87)
- FSRT: Facial Scene Representation Transformer for Face Reenactment from Factorized Appearance Head-pose and Facial Expression Features (CVPR 2024 · similarity 0.87)
- CapHuman: Capture Your Moments in Parallel Universes (CVPR 2024 · 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.
