Research map / Image generation / Editing and style transfer
A Closed-form Solution to Photorealistic Image Stylization
Yijun Li, Ming-Yu Liu, Xueting Li, Ming-Hsuan Yang, Jan Kautz
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 |
|---|---|---|---|
| replace WCT with PhotoWCT in stylization step (ablation) | Proposed method · User study, stylization | preference: 83.6% | Improved |
| replace WCT with PhotoWCT in stylization step (ablation) | Proposed method · User study, photorealism | preference: 83.2% | Improved |
| add the closed-form photorealistic smoothing step (vs guided filtering approx) (ablation) | Proposed method · User study, stylization | preference: 59.6% / 40.4 | Improved |
| add the closed-form smoothing step (vs guided filtering approx) (ablation) | Proposed method · User study, fewer artifacts | preference: 52.8% / 47.2 | Improved |
| use guided image filtering to approximate the smoothing step (ablation) | Proposed method · Run-time 1024x512 | seconds: 13.16 -> 0.64 | Improved |
| use guided image filtering approximation (ablation) | Proposed method · User study vs Luan et al., stylization | preference: approx 63.6% | Improved |
| decrease smoothing lambda (ablation) | Photorealistic smoothing · HED boundary ODS/OIS | photorealism | Mixed |
| PhotoWCT+smoothing instead of Luan et al. optimization | Photorealistic stylization · Run-time 1024x512 | seconds: 650.45 -> 13.16 | Improved |
| proposed method vs Luan et al. | Photorealistic stylization · User study, fewer artifacts | preference: 26.5% -> 73.5% | Improved |
| proposed method vs classical Pitie et al. | Photorealistic stylization · User study, stylization | preference: 44.8% -> 55.2% | Improved |
| proposed method vs Pitie et al. on artifacts | Photorealistic stylization · User study, fewer artifacts | preference: 48.8% / 51.2% | No clear effect |
Closest papers by meaning
- Ultrafast Photorealistic Style Transfer via Neural Architecture Search (AAAI 2020 · similarity 0.89)
- Styl3R: Instant 3D Stylized Reconstruction for Arbitrary Scenes and Styles (NeurIPS 2025 · similarity 0.87)
- ReGS: Reference-based Controllable Scene Stylization with Gaussian Splatting (NeurIPS 2024 · similarity 0.85)
- Multi-StyleGS: Stylized Gaussian Splatting with Multiple Styles (AAAI 2025 · similarity 0.85)
- Local Consistency Guidance: Personalized Stylization Method of Face Video (Student Abstract) (AAAI 2024 · similarity 0.84)
- StyO: Stylize Your Face in Only One-Shot (AAAI 2025 · similarity 0.84)
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.
