Research map / Image generation / Editing and style transfer
StyleCLIP: Text-Driven Manipulation of StyleGAN Imagery
Or Patashnik, Zongze Wu, Eli Shechtman, Daniel Cohen-Or, Dani Lischinski
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 a global direction instead of the latent optimizer for inference | StyleCLIP · Efficiency (GTX 1080Ti) | inference time: 98 sec -> 72 ms | Improved |
| use a trained latent mapper instead of the latent optimizer for inference | StyleCLIP · Efficiency (GTX 1080Ti) | inference time: 98 sec -> 75 ms | Improved |
| add prompt engineering (averaging a bank of sentence templates) | CLIP zero-shot classification · ImageNet zero-shot | accuracy: +3.5% | Improved |
| use three separate mapping networks (coarse, medium, fine) instead of a single mapping network (ablation) | Latent mapper · Donald Trump edit (qualitative) | edit quality | Improved |
| remove the fine-level mapper M^f for hairstyle edits | Latent mapper · Hairstyle edits (qualitative) | edit quality | Improved |
| add identity loss to the mapper objective (ablation) | Latent mapper · Identity preservation (qualitative) | identity preservation | Improved |
| use CLIP loss instead of identity loss to a single Beyonce image (ablation) | Latent mapper · Beyonce celeb edit (qualitative) | edit uniqueness | Improved |
| raise the disentanglement threshold beta (ablation) | Global direction · Grey hair edit (qualitative) | disentanglement vs edit strength | Mixed |
Closest papers by meaning
- Learning Input-agnostic Manipulation Directions in StyleGAN with Text Guidance (ICLR 2023 · similarity 0.92)
- Text-Guided Unsupervised Latent Transformation for Multi-Attribute Image Manipulation (CVPR 2023 · similarity 0.91)
- Learning to Manipulate Artistic Images (AAAI 2024 · similarity 0.89)
- TextCtrl: Diffusion-based Scene Text Editing with Prior Guidance Control (NeurIPS 2024 · similarity 0.88)
- The Devil is in the Details: StyleFeatureEditor for Detail-Rich StyleGAN Inversion and High Quality Image Editing (CVPR 2024 · similarity 0.88)
- Style Aligned Image Generation via Shared Attention (CVPR 2024 · 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.
