Research map / 3D vision / Face and shape
HeadGAN: One-Shot Neural Head Synthesis and Editing
Michail Christos Doukas, Stefanos Zafeiriou, Viktoriia Sharmanska
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 dense flow network F (warping reference features) (ablation) | HeadGAN · VoxCeleb ablation (half epochs) | FID: 63.3 -> 50.9 | Improved |
| add the dense flow network F (warping reference features) (ablation) | HeadGAN · VoxCeleb ablation (half epochs) | CSIM: 0.307 -> 0.716 | Improved |
| condition on 3D face representation instead of 2D landmark sketches (ablation) | HeadGAN · VoxCeleb ablation (half epochs) | FID: 55.7 -> 50.9 | Improved |
| condition on 3D face representation instead of 2D landmark sketches (ablation) | HeadGAN · VoxCeleb ablation (half epochs) | CSIM: 0.699 -> 0.716 | Improved |
| add audio features as input (ablation) | HeadGAN · VoxCeleb ablation (half epochs) | FID: 55.1 -> 50.9 | Improved |
| add audio features as input (ablation) | HeadGAN · BBC lipreading (25 word classes) | lipreading accuracy: 73% -> 82% | Improved |
| use absolute instead of relative key-point locations | FOMM · VoxCeleb test reenactment | CSIM: 0.765 -> 0.587 | Hurt |
| train on larger VoxCeleb2 (Bi-layer baseline) | Bi-layer Neural Avatars · VoxCeleb test reenactment | AU-H: 0.296 vs 0.326 | Mixed |
Later papers that compared with this method
2 later accepted papers report HeadGAN as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- Free-HeadGAN: Neural Talking Head Synthesis With Explicit Gaze Control (TPAMI 2023 · better on 10 of 11 shared results)
- Face2Faceρ: Real-Time High-Resolution One-Shot Face Reenactment (ECCV 2022 · better on 9 of 9 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
- Free-HeadGAN: Neural Talking Head Synthesis With Explicit Gaze Control (TPAMI 2023 · similarity 0.91)
- Enhancing Identity-Deformation Disentanglement in StyleGAN for One-Shot Face Video Re-Enactment (AAAI 2025 · similarity 0.88)
- VOODOO 3D: Volumetric Portrait Disentanglement For One-Shot 3D Head Reenactment (CVPR 2024 · similarity 0.87)
- CapHuman: Capture Your Moments in Parallel Universes (CVPR 2024 · similarity 0.87)
- Occlusion-Insensitive Talking Head Video Generation via Facelet Compensation (AAAI 2025 · similarity 0.87)
- ControlFace: Harnessing Facial Parametric Control for Face Rigging (CVPR 2025 · 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.
