AI Paper Atlas

Research map / Image generation / Text-to-image diffusion

StyleAdapter: A Unified Stylized Image Generation Model

Zhouxia Wang, Xintao Wang, Liangbin Xie, Zhongang Qi, Ying Shan, Wenping Wang, Ping Luo

IJCV 2024 · Image generation › Text-to-image diffusion · approach: Layout-to-image control · 92 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
add Two-Path Cross-Attention (TPCA) (ablation)Vanilla StyleAdapter · Single-referenceText-Sim: 0.1263 -> 0.2089Improved
add TPCA (ablation)Vanilla StyleAdapter · Single-referenceStyle-Sim: 0.9362 -> 0.8963Hurt
add TPCA (ablation)Vanilla StyleAdapter · Single-referenceFID: 186.17 -> 145.37Improved
remove E_cls (semantic suppression) from SSVM (ablation)TPCA · Single-referenceText-Sim: 0.2089 -> 0.2109Improved
add patch shuffling with E_pos (ablation)TPCA - E_cls · Single-referenceText-Sim: 0.2109 -> 0.2435Improved
add patch shuffling with E_pos (ablation)TPCA - E_cls · Single-referenceStyle-Sim: 0.8921 -> 0.8645Hurt
add TPCA to SSVM (vs VSA + SSVM) (ablation)StyleAdapter · Multi-referenceText-Sim: 0.2411 -> 0.2448Improved
add TPCA to SSVM (vs VSA + SSVM) (ablation)StyleAdapter · Multi-referenceFID: 160.82 -> 140.97Improved
use learnable embedding instead of averaging reference features (ablation)StyleEmb · Multi-referenceText-Sim: 0.2161 -> 0.2448Improved
use learnable embedding instead of concatenation (ablation)StyleEmb · Multi-referenceText-Sim: 0.2094 -> 0.2448Improved
use learnable embedding instead of concatenation (ablation)StyleEmb · Multi-referenceStyle-Sim: 0.8959 -> 0.9031Improved
shuffle via E_pos instead of raw image (ablation)SSVM · Single-referenceStyle-Sim: 0.8971 -> 0.9031Improved

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