Research map / Image generation / Editing and style transfer
Talk-To-Edit: Fine-Grained Facial Editing via Dialog
Yuming Jiang, Ziqi Huang, Xingang Pan, Chen Change Loy, Ziwei Liu
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 per-identity location-specific curved trajectories instead of a single straight direction (ablation) | Talk-to-Edit · Bangs/Eyeglasses/Beard/Smiling/Young editing | identity and attribute preservation: Bangs 0.7621 / 0.7491 -> 0.6047 / 0.3660 | Improved |
| use different directions for different identities (vs switching shared class-wise SVM directions) (ablation) | Talk-to-Edit · Bangs/Eyeglasses/Beard/Smiling/Young editing | identity and attribute preservation: Bangs 0.7262 / 0.5387 -> 0.6047 / 0.3660 | Improved |
| use semantic field direction conditioned on current location vs fixed identity-specific direction (ablation) | Talk-to-Edit · Smiling editing | identity and attribute preservation: 0.6639 / 0.5092 -> 0.6434 / 0.5028 | Improved |
| edit in W+ space instead of Z space | Talk-to-Edit · Smiling editing | identity and attribute preservation: 0.6434 / 0.5028 -> 0.4580 / 0.3573 | Improved |
| edit in W+ space instead of Z space | Talk-to-Edit · Eyeglasses editing | identity score: 0.6229 -> 0.6670 | Hurt |
| edit in W+ space instead of Z space | Talk-to-Edit · Young editing | identity and attribute preservation: 0.6309 / 0.4814 -> 0.6234 / 0.2731 | Improved |
| continue editing along the field with more attribute class changes | Talk-to-Edit · Eyeglasses (1024x1024, 100 latent codes) | cosine similarity to initial direction: cosine similarity decreases as attribute class change increases | 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 5 of 5 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
- Talk-to-Edit: Fine-Grained 2D and 3D Facial Editing via Dialog (TPAMI 2023 · similarity 0.99)
- SDGAN: Disentangling Semantic Manipulation for Facial Attribute Editing (AAAI 2024 · similarity 0.90)
- L2M-GAN: Learning To Manipulate Latent Space Semantics for Facial Attribute Editing (CVPR 2021 · similarity 0.89)
- Semantic Component Decomposition for Face Attribute Manipulation (CVPR 2019 · similarity 0.88)
- DreamSalon: A Staged Diffusion Framework for Preserving Identity-Context in Editable Face Generation (CVPR 2024 · similarity 0.87)
- CA-Edit: Causality-Aware Condition Adapter for High-Fidelity Local Facial Attribute Editing (AAAI 2025 · similarity 0.86)
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.
