AI Paper Atlas

Research map / Image generation / Editing and style transfer

MAUGen: A Unified Diffusion Approach for Multi-Identity Facial Expression and AU Label Generation

Xiangdong Li, Ye Lou, Ao Gao, Wei Zhang, Siyang Song

AAAI 2026 · Image generation › Editing and style transfer · approach: Diffusion face generation and editing · 1 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 identity features without decoupling (ablation)MAUGen · DISFA ablationF1 (%): 92.6 -> 88.5Hurt
add Identity Decoupling Modules (modulation mask + residual filtering) (ablation)MAUGen · DISFA ablationF1 (%): 88.5 -> 93.6Improved
add Identity Decoupling Modules (ablation)MAUGen · DISFA ablationFID: 6.0795 -> 5.9120Improved
add AU-guided mutual conditioning (ablation)MAUGen · DISFA ablationCLIP score: 0.2769 -> 0.2817Improved
add language-based query initialization (ablation)MAUGen · DISFA ablationF1 (%): to 95.7Improved
add LGFE (ablation)MAUGen · DISFA ablationF1 (%): 95.7 -> 96.2Improved
add pseudo-label image-label alignment loss with cross-detector agreement (ablation)MAUGen · DISFA ablationF1 (%): 96.2 -> 96.7Improved
train AU detectors on real plus synthetic MIFA dataFMAE (ViT-B) · DISFA test (real)F1 (%): 63.5 -> 70.9Improved
train AU detectors on real plus synthetic MIFA dataGraphAU (Swin-B) · DISFA test (real)F1 (%): 62.4 -> 64.1Improved
train on MIFA instead of DISFAGraphAU (Res50) · AffectNetF1 (%): 53.21 -> 56.01Improved
train on MIFA instead of DISFAFMAE (Vit-H) · AffectNetF1 (%): 70.58 -> 69.49Hurt
MAUGen vs baselines on image-text alignment (CLIP)MAUGen · DISFACLIP score: 0.295 vs StyleCLIP 0.303Mixed

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