AI Paper Atlas

Research map / Image generation / Editing and style transfer

A Training-Free Style-Personalization via SVD-Based Feature Decomposition

Kyoungmin Lee, Jihun Park, Jongmin Gim, Wonhyeok Choi, Kyumin Hwang, Jaeyeul Kim, Sunghoon Im

CVPR 2026 · Image generation › Editing and style transfer · approach: Arbitrary neural style transfer · 0 citations (OpenAlex)

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 changedWhereMeasuredEffect
replace dominant style feature directly (REP) instead of blending the SVD principal component (ablation)Infinity-2B · Parti-190S_txt: 0.348 -> 0.279Hurt
replace dominant style feature directly (REP) (ablation)Infinity-2B · Parti-190S_img: 0.559 -> 0.696Improved
add Principal Feature Blending (SVD-based) (ablation)Infinity-2B · Parti-190S_img: 0.559 -> 0.631Improved
add Principal Feature Blending (SVD-based) (ablation)Infinity-2B · Parti-190S_txt: 0.348 -> 0.321Hurt
use PFB instead of direct replacement (ablation)Infinity-2B · Parti-190S_txt: 0.279 -> 0.321Improved
add Structural Attention Correction on top of PFB (ablation)Infinity-2B · Parti-190S_txt: 0.321 -> 0.334Improved
add Structural Attention Correction on top of PFB (ablation)Infinity-2B · Parti-190S_img: 0.631 -> 0.630No clear effect
add SAC to PFB (ablation)Infinity-2B · Parti-190Harmonic score: 0.426 -> 0.437Improved
compare to training-free IP-Adapter and StyleAlignedOurs · Parti-190S_txt: 0.302/0.315 -> 0.334Improved
compare inference time to tuning-based methodsOurs · Parti-190Inference time (s): 3.58 vs 10.13 - 698.98Improved
compare to StyleAligned on harmonic scoreOurs · Parti-190Harmonic score: 0.437 vs 0.438No clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most 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

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy