AI Paper Atlas

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

Disentangling to Re-couple: Resolving the Similarity-Controllability Paradox in Subject-Driven Text-to-Image Generation

Shuang Li, Chao Deng, Hang Chen, Liqun Liu, Zhenyu Hu, Te Cao, Mengge Xue, Yuan Chen, Peng Shu, Huan Yu, Jie Jiang

CVPR 2026 · Image generation › Text-to-image diffusion · approach: Personalized and region-controlled generation · 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 textual-visual decoupling (ablation)DisCo · DreamBenchCLIP-I: 0.915 -> 0.928Improved
add textual-visual decoupling (ablation)DisCo · DreamBenchCLIP-T: 0.319 -> 0.329Improved
add textual-visual decoupling (ablation)DisCo · DreamBenchImageReward: 1.237 -> 1.339Improved
add GRPO re-coupling stage (ablation)DisCo · DreamBenchImageReward: 1.189 -> 1.339Improved
add GRPO re-coupling stage (ablation)DisCo · DreamBenchCLIP-I: 0.922 -> 0.928Improved
add GRPO re-coupling stage (ablation)DisCo · DreamBenchCLIP-T: 0.319 -> 0.329Improved
use trained pairwise VLM reward instead of CLIP reward (ablation)DisCo GRPO · DreamBenchCLIP-I: 0.898 -> 0.928Improved
use trained VLM reward instead of ImageReward reward (ablation)DisCo GRPO · DreamBenchCLIP-I: 0.914 -> 0.928Improved
use trained VLM reward instead of ImageReward reward (ablation)DisCo GRPO · DreamBenchImageReward: 1.404 -> 1.339Hurt
use trained VLM reward instead of untuned pretrained VLM reward (ablation)DisCo GRPO · DreamBenchImageReward: 1.189 -> 1.339Improved
apply naive prompt decoupling to DreamODreamO · DreamBenchImageReward: 1.186 -> 0.661Hurt
apply naive prompt decoupling to DreamODreamO · DreamBenchDINO-I: 0.813 -> 0.771Hurt

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