Research map / Image generation / Editing and style transfer
Rethinking Style Transfer: From Pixels to Parameterized Brushstrokes
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 |
|---|---|---|---|
| Optimize parameterized brushstrokes instead of pixels | Gatys-style artistic style transfer · 8 styles | mean deception score: 0.389 -> 0.588 | Helped |
| Optimize parameterized brushstrokes instead of pixels | Gatys-style artistic style transfer · 8 styles | mean human deception rate: 0.139 -> 0.268 | Helped |
| Optimize parameterized brushstrokes instead of pixels | artistic style transfer · 8 styles | mean deception score: 0.451 -> 0.588 | Helped |
| Optimize parameterized brushstrokes instead of pixels | artistic style transfer · 8 styles | mean human deception rate: 0.146 -> 0.268 | Helped |
| Use the explicit differentiable renderer instead of the trained renderer (ablation) | brushstroke rendering · brushstroke generation | memory use | Helped |
| Use the explicit differentiable renderer instead of the trained renderer (ablation) | brushstroke rendering · brushstroke generation | rendering speed | Mixed |
| Use the explicit differentiable renderer instead of Learning to Paint | image reconstruction · CelebA, 200 strokes | MSE: 20% lower | Helped |
| Use the explicit differentiable renderer instead of Learning to Paint | image reconstruction · CelebA, 1000 strokes | MSE: 49% lower | Helped |
| Increase the number of brushstrokes from 1K to 20K (ablation) | explicit brushstroke renderer · renderer runtime | speed: 1.17 $\pm$ 0.01 -> 0.93 $\pm$ 0.01 | Hurt |
Closest papers by meaning
- Stylized Neural Painting (CVPR 2021 · similarity 0.90)
- Neural 3D Strokes: Creating Stylized 3D Scenes with Vectorized 3D Strokes (CVPR 2024 · similarity 0.89)
- Adjustable Real-time Style Transfer (ICLR 2020 · similarity 0.89)
- MetaStyle: Three-Way Trade-off among Speed, Flexibility, and Quality in Neural Style Transfer (AAAI 2019 · similarity 0.88)
- ArtBank: Artistic Style Transfer with Pre-trained Diffusion Model and Implicit Style Prompt Bank (AAAI 2024 · similarity 0.87)
- Z*: Zero-shot Style Transfer via Attention Reweighting (CVPR 2024 · similarity 0.87)
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.