Research map / Image generation / Text-to-image diffusion
IntroStyle: Training-Free Introspective Style Attribution using Diffusion Features
Anand Kumar, Jiteng Mu, Nuno Vasconcelos
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 up block index 1 instead of 0 (ablation) | IntroStyle · WikiArt | mAP@10: 0.830 -> 0.852 | Improved |
| use up block index 1 instead of 3 (ablation) | IntroStyle · WikiArt | mAP@10: 0.836 -> 0.852 | Improved |
| use up block index 2 instead of 1 (ablation) | IntroStyle · WikiArt | mAP@10: 0.852 -> 0.852 | No clear effect |
| use W2 distance instead of Gram matrix (ablation) | IntroStyle · WikiArt | mAP@10: 0.835 -> 0.852 | Improved |
| use W2 distance instead of Euclidean L2 (ablation) | IntroStyle · WikiArt | mAP@10: 0.852 -> 0.852 | No clear effect |
| use W2 distance instead of JSD (ablation) | IntroStyle · WikiArt | mAP@10: 0.849 -> 0.852 | No clear effect |
| use W2 distance instead of L2 (ablation) | IntroStyle · DomainNet | mAP@10: 0.944 -> 0.949 | Improved |
| use up block index 1 instead of 0 (ablation) | IntroStyle · DomainNet | mAP@10: 0.925 -> 0.949 | Improved |
| increase timestep beyond about 400 (ablation) | IntroStyle · WikiArt | mAP@10 | Hurt |
| use IntroStyle instead of CSD ViT-L | IntroStyle · WikiArt | mAP@10: 0.538 -> 0.852 | Improved |
| use IntroStyle instead of CSD ViT-L | IntroStyle · DomainNet | mAP@10: 0.811 -> 0.949 | Improved |
| use IntroStyle instead of CSD ViT-L | IntroStyle · WikiArt | Recall@100: 0.956 -> 0.941 | No clear effect |
Closest papers by meaning
- StyleSSP: Sampling StartPoint Enhancement for Training-free Diffusion-based Method for Style Transfer (CVPR 2025 · similarity 0.88)
- Stylebreeder: Exploring and Democratizing Artistic Styles through Text-to-Image Models (NeurIPS 2024 · similarity 0.87)
- OmniStyle: Filtering High Quality Style Transfer Data at Scale (CVPR 2025 · similarity 0.87)
- SigStyle: Signature Style Transfer via Personalized Text-to-Image Models (AAAI 2025 · similarity 0.86)
- Style Injection in Diffusion: A Training-free Approach for Adapting Large-scale Diffusion Models for Style Transfer (CVPR 2024 · similarity 0.86)
- CSGO: Content-Style Composition in Text-to-Image Generation (NeurIPS 2025 · similarity 0.86)
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.
