Research map / Image generation / Editing and style transfer
Multi-Directional Subspace Editing in Style-Space
Chen Naveh
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 orthogonality loss between attribute subspaces (ablation) | MDSE · attribute correlation (Smile avg) | avg abs correlation: 0.242 -> 0.072 | Improved |
| add orthogonality loss between attribute subspaces (ablation) | MDSE · identity preservation, all edits | Euclidean distance: 0.536 -> 0.467 | Improved |
| use multi-directional subspace editing instead of single SVM direction (ablation) | MDSE · diversity of edits | LPIPS: 0.229 -> 0.321 | Improved |
| use multi-directional subspace editing instead of single SVM direction (ablation) | MDSE · diversity of edits | FID: 26.926 -> 23.648 | Improved |
| orthogonal subspaces vs StyleFlow for multi-attribute edits | MDSE · attribute correlation (Age avg) | avg abs correlation: 0.366 -> 0.219 | Improved |
| orthogonal subspaces vs StyleFlow for single smile edit identity | MDSE · identity preservation, smile | cosine similarity: 0.989 -> 0.991 | No clear effect |
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 3 of 3 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
- Image-to-Image Translation With Disentangled Latent Vectors for Face Editing (TPAMI 2023 · similarity 0.91)
- Adaptive Nonlinear Latent Transformation for Conditional Face Editing (ICCV 2023 · similarity 0.90)
- L2M-GAN: Learning To Manipulate Latent Space Semantics for Facial Attribute Editing (CVPR 2021 · similarity 0.89)
- StyleSpace Analysis: Disentangled Controls for StyleGAN Image Generation (CVPR 2021 · similarity 0.89)
- SDGAN: Disentangling Semantic Manipulation for Facial Attribute Editing (AAAI 2024 · similarity 0.88)
- A Latent Transformer for Disentangled Face Editing in Images and Videos (ICCV 2021 · 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.
