Research map / Image generation / GANs
Disentangled Lifespan Face Synthesis
Sen He, Wentong Liao, Michael Ying Yang, Yi-Zhe Song, Bodo Rosenhahn, Tao Xiang
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 disentangled shape/texture/identity representation instead of an entangled latent feature (ablation) | Lifespan face synthesis model · FFHQ qualitative (Fig.) | image quality, shape deformation, texture transformation, reconfiguration | Improved |
| add shape regularization loss L_s (small shape change between adult age groups) (ablation) | Lifespan face synthesis model · FFHQ qualitative (Fig.) | texture leakage into shape features | Improved |
| add texture transformation on top of shape transformation (ablation) | Lifespan face synthesis model · FFHQ qualitative (Fig.) | texture change in older age groups | Improved |
| transform shape features f_s alone with texture f_t fixed (ablation) | Lifespan face synthesis model · FFHQ qualitative (Fig.) | shape vs texture disentanglement | Improved |
| compare to LATS on identity preservation (user study) | Ours vs LATS · FFHQ lifespan user study (AMT) | identity preservation: 3.18 -> 3.07 | Hurt |
| compare to LATS on shape transformation (user study) | Ours vs LATS · FFHQ lifespan user study (AMT) | shape transformation: 2.89 -> 3.18 | Improved |
| compare to LATS on age accuracy | Ours vs LATS · FFHQ lifespan user study (AMT) | age accuracy: 60.0% -> 65.6% | Improved |
| compare to LATS on reconfiguration (LPIPS, lower better) | Ours vs LATS · FFHQ lifespan automatic evaluation | Reconfig LPIPS: 0.11 -> 0.07 | Improved |
| compare to LATS on identity preservation (VGG-face) | Ours vs LATS · FFHQ lifespan automatic evaluation | ID: 96.68% -> 96.47% | No clear effect |
Closest papers by meaning
- Lifespan Age Transformation Synthesis (ECCV 2020 · similarity 0.91)
- Age Progression and Regression with Spatial Attention Modules (AAAI 2020 · similarity 0.90)
- TimeBooth: Disentangled Facial Invariant Representation for Diverse and Personalized Face Aging (ICCV 2025 · similarity 0.90)
- Continuous Face Aging via Self-Estimated Residual Age Embedding (CVPR 2021 · similarity 0.90)
- Face Aging With Identity-Preserved Conditional Generative Adversarial Networks (CVPR 2018 · similarity 0.89)
- From Cradle to Cane: A Two-Pass Framework for High-Fidelity Lifespan Face Aging (NeurIPS 2025 · similarity 0.88)
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.
