Research map / Image generation / GANs
HairMapper: Removing Hair From Portraits Using GANs
Yiqian Wu, Yong-Liang Yang, Xiaogang Jin
Check 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 HairMapper rather than InterFaceGAN for hair removal | StyleGAN2-ada latent codes · 500 randomly sampled latent codes | identity loss (Lid): 0.5933 -> 0.3451 | Improved |
| Use HairMapper rather than StyleSpace for hair removal | StyleGAN2-ada latent codes · 500 randomly sampled latent codes | identity loss (Lid): 0.3382 -> 0.3451 | Hurt |
| Use HairMapper rather than StyleClip for hair removal | StyleGAN2-ada latent codes · 500 randomly sampled latent codes | identity loss (Lid): 0.3986 -> 0.3451 | Improved |
| Use HairMapper rather than StyleFlow for hair removal | StyleGAN2-ada latent codes · 500 randomly sampled latent codes | identity loss (Lid): 0.4257 -> 0.3451 | Improved |
| Use HairMapper for hair removal instead of the four compared baselines | hair-removal outputs · Hair-removal preference user study | chosen as best: 66.86% vs. 8.43%, 6.14%, 9.86%, and 8.71% | Improved |
| Train the gender boundary from StyleFlow gender-transition pairs rather than random codes with gender scores (ablation) | female-to-male latent-code transformation · Figure 7 comparison | face-shape and pose preservation | Improved |
| Initialize hair-removal diffusion with the edited bald code rather than the original haired code (ablation) | male hair-removal optimization · Figure 8 comparison | hair removal quality | Improved |
| Dilate and blur the hair mask before Poisson blending (ablation) | HairMapper image blending · portrait hair removal | shadow and blending artifacts | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- InterFaceGAN — InterFaceGAN: Interpreting the Disentangled Face Representation Learned by GANs (TPAMI 2020) · better on 2 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
- GAN with Multivariate Disentangling for Controllable Hair Editing (ECCV 2022 · similarity 0.88)
- Real-Time Hair Rendering using Sequential Adversarial Networks (ECCV 2018 · similarity 0.88)
- StyleGAN Salon: Multi-View Latent Optimization for Pose-Invariant Hairstyle Transfer (CVPR 2023 · similarity 0.87)
- HairNet: Single-View Hair Reconstruction using Convolutional Neural Networks (ECCV 2018 · similarity 0.86)
- Feature Unlearning for Pre-trained GANs and VAEs (AAAI 2024 · similarity 0.85)
- DiffLocks: Generating 3D Hair from a Single Image using Diffusion Models (CVPR 2025 · similarity 0.84)
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.
