AI Paper Atlas

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

FocusDPO: Dynamic Preference Optimization for Multi-Subject Personalized Image Generation via Adaptive Focus

Qiaoqiao Jin, Siming Fu, Dong She, Weinan Jia, Hualiang Wang, Mu Liu, Jidong Jiang

AAAI 2026 · Image generation › Text-to-image diffusion · approach: Multi-subject personalization · 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 FocusDPO training to UNOUNO (FLUX.1 dev) · DreamBench (single-subject)DINO: 0.760 -> 0.802Improved
add FocusDPO training to UNOUNO (FLUX.1 dev) · DreamBench (multi-subject)DINO: 0.542 -> 0.570Improved
add FocusDPO training to IP-Adapter-PlusIP-Adapter-Plus (SDXL) · DreamBench (single-subject)DINO: 0.692 -> 0.751Improved
use FocusDPO instead of PatchDPOIP-Adapter-Plus (SDXL) · DreamBench (single-subject)DINO: 0.727 -> 0.751Improved
use FocusDPO instead of PatchDPOIP-Adapter-Plus (SDXL) · DreamBench (single-subject)CLIP-T: 0.292 -> 0.303Improved
use FocusDPO instead of standard DPO (ablation)UNO (FLUX.1 dev) · DreamBench (single-subject)DINO: 0.764 -> 0.802Improved
use FocusDPO instead of SFT (ablation)UNO (FLUX.1 dev) · DreamBench (multi-subject)DINO: 0.542 -> 0.570Improved
use FocusDPO instead of standard DPO (ablation)UNO (FLUX.1 dev) · DreamBench (multi-subject)CLIP-I: 0.731 -> 0.739Improved
add SFT on DIP pairs (ablation)UNO (FLUX.1 dev) · DreamBench (single-subject)CLIP-T: 0.304 -> 0.299No clear effect
add standard DPO on DIP pairs (ablation)UNO (FLUX.1 dev) · DreamBench (multi-subject)DINO: 0.542 -> 0.545No clear effect
add standard DPO on DIP pairs (ablation)IP-Adapter-Plus (SDXL) · DreamBench (single-subject)CLIP-T: 0.281 -> 0.276Hurt
add FocusDPO to IP-Adapter-Plus (text alignment)IP-Adapter-Plus (SDXL) · DreamBench (single-subject)CLIP-T: 0.281 -> 0.303Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 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