RECORD
TRACE: Time-Adaptive Residual Attention Control with Content-Style Decomposition for Training-Free Diffusion Style Transfer
Computer Vision and Pattern Recognition · Submitted 4 Oct 2026
abstract compiled · 2026-10-06 · no language-model review
The authors' main claim Experiments show that TRACE achieves a favorable trade-off between stylization and preservation.
The compiled record
The paper's own sentences, arranged. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.
Open the compiled record
The abstract, sentence by sentence
Reference-guided style transfer aims to preserve the semantic structure of a content image while transferring the visual appearance of a style reference. Recent diffusion-based methods achieve impressive stylization quality by exploiting strong pretrained generative priors. However, training-free approaches still face a difficult trade-off among style fidelity, content preservation, and content leakage. Direct style injection may unintentionally transfer semantic content from the style image, while fixed guidance schedules often ignore the time- and state-dependent nature of diffusion sampling. To address these limitations, we propose TRACE, a training-free diffusion style transfer framework with Time-adaptive Residual Attention Control and Content-Style Decomposition. TRACE first performs offline CLIP-based subspace analysis to separate content and style directions from paired data. During inference, it removes content-related components from the style reference and style-related components from the content reference to reduce leakage. It then injects style information through residual cross-attention and applies uncertainty-aware guidance to adapt the guidance signal at each denoising step. Experiments show that TRACE achieves a favorable trade-off between stylization and preservation. Compared with optimal-control-based baselines, TRACE substantially improves style fidelity (+17.28 CSD and +34.10 SRA). While, compared with stylization methods, it better preserves content structure (+12.80 DINO, +5.52 CLIP-I, and -8.19 LPIPS) and reduces directional semantic leakage by 29.5% in DCL. Our code is publicly available at https://github.com/pixelchemy-research/TRACE.
background contribution method result interpretation main claim
| # | sentence | role | the authors' own work | states the main conclusion | strength |
|---|---|---|---|---|---|
| s0 | Reference-guided style transfer aims to preserve the semantic structure of a content image while transferring the visual appearance of a style reference. | background 1.00 | 0.04 | 0.02 | asserted 0.99 |
| s1 | Recent diffusion-based methods achieve impressive stylization quality by exploiting strong pretrained generative priors. | background 1.00 | 0.05 | 0.02 | asserted 1.00 |
| s2 | However, training-free approaches still face a difficult trade-off among style fidelity, content preservation, and content leakage. | gap 0.70 | 0.08 | 0.04 | asserted 1.00 |
| s3 | Direct style injection may unintentionally transfer semantic content from the style image, while fixed guidance schedules often ignore the time- and state-dependent nature of diffusion sampling. | gap 0.59 | 0.06 | 0.02 | hedged possibility 0.69 |
| s4 | To address these limitations, we propose TRACE, a training-free diffusion style transfer framework with Time-adaptive Residual Attention Control and Content-Style Decomposition. | contribution 0.99 | 0.80 | 0.07 | asserted 0.99 |
| s5 | TRACE first performs offline CLIP-based subspace analysis to separate content and style directions from paired data. | method 0.99 | 0.87 | 0.03 | asserted 1.00 |
| s6 | During inference, it removes content-related components from the style reference and style-related components from the content reference to reduce leakage. | method 0.99 | 0.86 | 0.04 | asserted 1.00 |
| s7 | It then injects style information through residual cross-attention and applies uncertainty-aware guidance to adapt the guidance signal at each denoising step. | method 1.00 | 0.85 | 0.05 | asserted 1.00 |
| s8 | Experiments show that TRACE achieves a favorable trade-off between stylization and preservation. | result 0.94 | 0.95 | 0.69 | asserted 1.00 |
| s9 | Compared with optimal-control-based baselines, TRACE substantially improves style fidelity (+17.28 CSD and +34.10 SRA). | result 1.00 | 0.95 | 0.22 | asserted 1.00 |
| s10 | While, compared with stylization methods, it better preserves content structure (+12.80 DINO, +5.52 CLIP-I, and -8.19 LPIPS) and reduces directional semantic leakage by 29.5% in DCL. | result 1.00 | 0.94 | 0.27 | asserted 1.00 |
| s11 | Our code is publicly available at https://github.com/pixelchemy-research/TRACE. | contribution 0.66 | 0.55 | 0.02 | asserted 0.95 |
What was examined, and what was not
- arXiv listing abstract v1
- full text
- figures and tables
- references and citing sentences
- relations to other papers
- P abstract
- S sections
- T body sentences
- C citations
- R cross-paper relations
The paper
Original abstract
Reference-guided style transfer aims to preserve the semantic structure of a content image while transferring the visual appearance of a style reference. Recent diffusion-based methods achieve impressive stylization quality by exploiting strong pretrained generative priors. However, training-free approaches still face a difficult trade-off among style fidelity, content preservation, and content leakage. Direct style injection may unintentionally transfer semantic content from the style image, while fixed guidance schedules often ignore the time- and state-dependent nature of diffusion sampling. To address these limitations, we propose TRACE, a training-free diffusion style transfer framework with Time-adaptive Residual Attention Control and Content-Style Decomposition. TRACE first performs offline CLIP-based subspace analysis to separate content and style directions from paired data. During inference, it removes content-related components from the style reference and style-related components from the content reference to reduce leakage. It then injects style information through residual cross-attention and applies uncertainty-aware guidance to adapt the guidance signal at each denoising step. Experiments show that TRACE achieves a favorable trade-off between stylization and preservation. Compared with optimal-control-based baselines, TRACE substantially improves style fidelity (+17.28 CSD and +34.10 SRA). While, compared with stylization methods, it better preserves content structure (+12.80 DINO, +5.52 CLIP-I, and -8.19 LPIPS) and reduces directional semantic leakage by 29.5% in DCL. Our code is publicly available at https://github.com/pixelchemy-research/TRACE.
Human review
Discussion
Know something about this paper? Leave a comment, or highlight any passage above to comment on it.