Research map / Image generation / Editing and style transfer
Multimodal Style Transfer via Graph Cuts
Yulun Zhang, Chen Fang, Yilin Wang, Zhaowen Wang, Zhe Lin, Yun Fu, Jimei Yang
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 cosine distance instead of Euclidean distance for graph-cut style matching (ablation) | MST · Qualitative style transfer | visual quality (wash-out artifacts) | Improved |
| add the smoothness term (lambda=0.1) rather than lambda=0 (ablation) | MST · Portrait stylization (qualitative) | visual quality (face smoothness): lambda 0 -> 0.1 | Improved |
| increase smoothness weight lambda from 0.1 to 1 (ablation) | MST · Portrait stylization (qualitative) | style diversity: lambda 0.1 -> 1 | Hurt |
| use whitening and coloring instead of AdaIN statistics for feature transform (K=1) (ablation) | MST-1 · Qualitative style transfer | visual quality (stroke artifacts) | Improved |
| increase style cluster number K from 1 to larger values (ablation) | MST · Qualitative style transfer | match of style patterns to content | Mixed |
| increase style cluster number K from 1 to 5 (ablation) | MST · Running time (100 pairs, 512x512) | time (s): 0.20 -> 2.27 | Hurt |
| add multimodal style clustering and graph-cut matching to AdaIN | AdaIN · Qualitative style transfer | content structure preservation | Improved |
| use MST (K=3) instead of Gatys et al. optimization | style transfer · User study (2,000 votes) | vote share (%): 21.41 -> 33.45 | Improved |
Closest papers by meaning
- Multimodal Transfer: A Hierarchical Deep Convolutional Neural Network for Fast Artistic Style Transfer (CVPR 2017 · similarity 0.91)
- Universal Style Transfer via Feature Transforms (NeurIPS 2017 · similarity 0.91)
- Learning Graph Neural Networks for Image Style Transfer (ECCV 2022 · similarity 0.91)
- In the Light of Feature Distributions: Moment Matching for Neural Style Transfer (CVPR 2021 · similarity 0.90)
- SigStyle: Signature Style Transfer via Personalized Text-to-Image Models (AAAI 2025 · similarity 0.88)
- Semantix: An Energy-guided Sampler for Semantic Style Transfer (ICLR 2025 · similarity 0.88)
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.
