Research map / Image generation / Editing and style transfer
Balanced Image Stylization with Style Matching Score
Yuxin Jiang, Liming Jiang, Shuai Yang, Jia-Wei Liu, Ivor W. Tsang, Mike Zheng Shou
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 semantic-aware gradient refinement R (ablation) | SMS · Ghibli style (PIE-bench) | ArtFID: 27.514 -> 18.686 (w/o R 24.132) | Improved |
| add progressive spectrum regularization L_freq (ablation) | SMS · Ghibli style (PIE-bench) | ArtFID: 27.514 -> 18.686 | Improved |
| add both L_freq and R (ablation) | SMS · Ghibli style (PIE-bench) | LPIPS: 0.703 -> 0.326 | Improved |
| use adaptive narrowing sampling instead of random t (ablation) | SMS · Ghibli style (PIE-bench) | FID: 22.766 -> 13.089 | Improved |
| use adaptive narrowing sampling instead of naive linear t (ablation) | SMS · Ghibli style (PIE-bench) | ArtFID: 23.524 -> 18.686 | Improved |
| use SMS instead of Style-LoRA with ControlNet | SMS · Ghibli style (PIE-bench) | LPIPS: 0.438 -> 0.326 | Improved |
| use SMS instead of DDS | SMS · Oil painting style (PIE-bench) | ArtFID: 37.864 -> 27.055 | Improved |
| use SMS style LoRA prior vs Style-LoRA on style metric | SMS · Ghibli style (PIE-bench) | PickScore: 2.067 -> 1.487 | Hurt |
| distill SMS into feed-forward generator instead of DDS | SMS · Feed-forward Ghibli (LHQ) | LPIPS: 0.321 -> 0.268 | Improved |
| use SMS instead of Scenimefy (GAN) in feed-forward | SMS · Feed-forward Ghibli (LHQ) | FID: 12.054 -> 12.472 | Hurt |
| increase loss weight lambda (ablation) | SMS · qualitative | content vs style | Mixed |
| replace progressive spectrum with fixed threshold 0.3 / spatial MSE / E-LatentLPIPS (ablation) | SMS · qualitative | visual quality | Hurt |
Closest papers by meaning
- DEADiff: An Efficient Stylization Diffusion Model with Disentangled Representations (CVPR 2024 · similarity 0.88)
- SigStyle: Signature Style Transfer via Personalized Text-to-Image Models (AAAI 2025 · similarity 0.88)
- Conditional Balance: Improving Multi-Conditioning Trade-Offs in Image Generation (CVPR 2025 · similarity 0.88)
- OmniStyle: Filtering High Quality Style Transfer Data at Scale (CVPR 2025 · similarity 0.87)
- One-Shot Structure-Aware Stylized Image Synthesis (CVPR 2024 · similarity 0.87)
- Style Injection in Diffusion: A Training-free Approach for Adapting Large-scale Diffusion Models for Style Transfer (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.
