Research map / Image generation / Editing and style transfer
UnZipLoRA: Separating Content and Style from a Single Image
Chang Liu, Viraj Shah, Aiyu Cui, Svetlana Lazebnik
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 prompt separation (ablation) | DreamBooth-LoRA baseline · Subject decomposition user study | % preference: 91.67% | Improved |
| add prompt separation (ablation) | DreamBooth-LoRA baseline · Style decomposition user study | % preference: 12.35% | Hurt |
| add prompt separation (ablation) | DreamBooth-LoRA baseline · Combined recontextualization user study | % preference: 92.80% | Improved |
| add column separation (ablation) | M1 · Subject decomposition user study | % preference: 55.74% | Improved |
| add column separation (ablation) | M1 · Style decomposition user study | % preference: 39.51% | Hurt |
| add column separation (ablation) | M1 · Combined recontextualization user study | % preference: 93.64% | Improved |
| add block separation (ablation) | M2 · Subject decomposition user study | % preference: 55.36% | Improved |
| add block separation (ablation) | M2 · Style decomposition user study | % preference: 86.42% | Improved |
| add block separation (ablation) | M2 · Combined recontextualization user study | % preference: 61.90% | Improved |
| add block separation (ablation) | M2 · Content-style decomposition | Style-alignment (CSD): 0.216 -> 0.265 | Improved |
| add block separation (ablation) | M2 · Content-style decomposition | Style-alignment (CLIP-I): 0.407 -> 0.427 | Improved |
| add prompt separation (ablation) | DreamBooth-LoRA · Content-style decomposition | Style-alignment (CSD): 0.245 -> 0.217 | Hurt |
Later papers that compared with this method
3 later accepted papers report UnZipLoRA as a baseline in their results tables; 3 of them report a better number on most of the results they share with it.
- SplitFlux: Learning to Decouple Content and Style from a Single Image (CVPR 2026 · better on 5 of 6 shared results)
- AnyStyle: A Single LoRA is Sufficient for Image-Guided Style Transfer (ECCV 2026 · better on 5 of 5 shared results)
- RefDiT: Local Attribute Guidance in Reference-Based Image Generation (ECCV 2026 · better on 3 of 3 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
- Implicit Style-Content Separation using B-LoRA (ECCV 2024 · similarity 0.87)
- K-LoRA: Unlocking Training-Free Fusion of Any Subject and Style LoRAs (CVPR 2025 · similarity 0.85)
- StyleSSP: Sampling StartPoint Enhancement for Training-free Diffusion-based Method for Style Transfer (CVPR 2025 · similarity 0.83)
- UnlearnCanvas: Stylized Image Dataset for Enhanced Machine Unlearning Evaluation in Diffusion Models (NeurIPS 2024 · similarity 0.82)
- Cached Multi-Lora Composition for Multi-Concept Image Generation (ICLR 2025 · similarity 0.82)
- Less is More: Masking Elements in Image Condition Features Avoids Content Leakages in Style Transfer Diffusion Models (ICLR 2025 · similarity 0.82)
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.
