SafeStyle: Calibrated Style Residual Injection for Controllable Style-Leakage Trade-off in Diffusion Stylization

Researchers introduced SafeStyle, a training-free framework for calibrated style residual injection in frozen diffusion models that balances style fidelity against content leakage in reference-guided diffusion stylization. SafeStyle estimates style-supported and content-associated subspaces from compact calibration sets, transports purified style evidence over adaptive spatial granularity, and constrains its influence with an explicit residual-norm budget. Experiments across texture- and geometry-dominant styles report a DINO style similarity of 0.432 with competitive text alignment, and on a semantically disjoint leakage-stress benchmark a DINO style similarity of 0.474 with only 0.8% semantic leakage.

arXiv:2609.21242v1 Announce Type: new Abstract: Reference-guided diffusion stylization aims to transfer visual style from a reference image while preserving the semantics specified by a text prompt. However, image conditioning often entangles transferable style cues with reference-specific content, leading to an inherent trade-off: stronger conditioning improves style fidelity but increases content leakage, whereas aggressive suppression reduces leakage at the cost of style expression. This challenge is further complicated by the distinct spatial organization of texture- and geometry-dominant styles. To address these issues, we propose SafeStyle, a training-free framework for calibrated style residual injection in frozen diffusion models. SafeStyle first estimates style-supported and content-associated subspaces from compact calibration sets, preserving their informative overlap while suppressing useless content variations. It then transports the purified style evidence over adaptive spatial granularity and constrains its effective influence through an explicit residual-norm budget. Experiments across texture- and geometry-dominant styles show that SafeStyle achieves a DINO style similarity of 0.432 while maintaining competitive text alignment. On a semantically disjoint leakage-stress benchmark, it further achieves a DINO style similarity of 0.474 with only 0.8\% semantic leakage, demonstrating an effective balance between style fidelity and reference-content suppression.