Research map / Image generation / Editing and style transfer
L2M-GAN: Learning To Manipulate Latent Space Semantics for Facial Attribute Editing
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 L2M-GAN instead of the strongest baseline by FID (avg), PA-GAN | Facial attribute editing · CelebA-HQ test, Smiling | FID (avg): 21.0 -> 20.6 | Helped |
| Use L2M-GAN instead of the strongest baseline by attribute accuracy, InterfaceGAN* | Facial attribute editing · CelebA-HQ test, Smiling | Acc (att): 92.1% -> 93.1% | Helped |
| Add perceptual loss (ablation) | StarGAN v2 baseline · CelebA-HQ test, Smiling | FID (avg): 25.3 -> 22.5 | Helped |
| Add perceptual loss (ablation) | StarGAN v2 baseline · CelebA-HQ test, Smiling | Acc (att): 77.2% -> 78.7% | Helped |
| Add single-term inner-product orthogonality loss to the perceptual-loss model (ablation) | StarGAN v2 baseline · CelebA-HQ test, Smiling | Acc (att): 78.7% -> 88.3% | Helped |
| Enforce inner-product orthogonality before and after transformation (ablation) | StarGAN v2 with perceptual loss and single-term orthogonality · CelebA-HQ test, Smiling | Acc (att): 88.3% -> 92.0% | Helped |
| Use two-term L1 orthogonality instead of two-term inner-product orthogonality (ablation) | StarGAN v2 with perceptual loss and orthogonality loss · CelebA-HQ test, Smiling | Acc (att): 92.0% -> 93.1% | Helped |
| Add perceptual loss to two-term L1 orthogonality loss (ablation) | StarGAN v2 baseline with two-term L1 orthogonality loss · CelebA-HQ test, Smiling | Acc (att): 91.8% -> 93.1% | Helped |
| Add orthogonality loss to perceptual-loss baseline (ablation) | StarGAN v2 baseline · CelebA-HQ test, Smiling | FID (avg): 22.5 -> 21.7 | Helped |
Closest papers by meaning
- SDGAN: Disentangling Semantic Manipulation for Facial Attribute Editing (AAAI 2024 · similarity 0.92)
- Interpreting the Latent Space of GANs for Semantic Face Editing (CVPR 2020 · similarity 0.91)
- Semantic Component Decomposition for Face Attribute Manipulation (CVPR 2019 · similarity 0.90)
- The GAN That Warped: Semantic Attribute Editing With Unpaired Data (CVPR 2020 · similarity 0.89)
- Enjoy Your Editing: Controllable GANs for Image Editing via Latent Space Navigation (ICLR 2021 · similarity 0.89)
- STGAN: A Unified Selective Transfer Network for Arbitrary Image Attribute Editing (CVPR 2019 · similarity 0.89)
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.