AI Paper Atlas

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

CharaConsist: Fine-Grained Consistent Character Generation

Mengyu Wang, Henghui Ding, Jianing Peng, Yao Zhao, Yunpeng Chen, Yunchao Wei

ICCV 2025 · Image generation › Text-to-image diffusion · approach: Layout-to-image control · 8 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 point-tracking attention (ablation)FLUX.1 consistent generation · Background maintainingID Sim: 0.476 -> 0.572Improved
add adaptive token merge on top of point-tracking attention (ablation)CharaConsist · Background maintainingID Sim: 0.572 -> 0.645Improved
add foreground-background masking for background consistency (ablation)CharaConsist · Background maintainingCLIP-I-bg: 0.899 -> 0.916Improved
add foreground-background masking (ablation)CharaConsist · Background maintainingCLIP-I-fg: 0.883 -> 0.883No clear effect
add all three components to FLUX.1 (ablation)FLUX.1 · Background maintainingCLIP-I: 0.874 -> 0.910Improved
add the proposed modules (image quality cost) (ablation)FLUX.1 · Image qualityIQS: 0.989 -> 0.985Hurt
add CharaConsist to base model (consistency increment) (ablation)FLUX.1 · Background maintainingID Sim: +.171Improved
use CharaConsist instead of ConsiStoryFLUX.1 vs SDXL · Background maintainingCLIP-I-bg: 0.895 -> 0.916Improved
use CharaConsist instead of StoryDiffusionFLUX.1 vs RealVisXL4.0 · Background switchingID Sim: 0.563 -> 0.587Improved
use CharaConsist (training-free) instead of PhotoMaker (trained)FLUX.1 · Background maintainingID Sim: 0.730 -> 0.647Hurt
use CharaConsist instead of StoryDiffusionFLUX.1 vs RealVisXL4.0 · Background maintainingCLIP-T: 0.370 -> 0.356Hurt
use CharaConsist instead of StoryDiffusionFLUX.1 · Background switchingCLIP-I: 0.853 -> 0.840Hurt

Later papers that compared with this method

1 later accepted paper reports CharaConsist as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

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