Saved in:
Bibliographic Details
Main Authors: Gao, Zhanxin, Zhu, Beier, Yao, Liang, Yang, Jian, Tai, Ying
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2507.08396
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918316704006144
author Gao, Zhanxin
Zhu, Beier
Yao, Liang
Yang, Jian
Tai, Ying
author_facet Gao, Zhanxin
Zhu, Beier
Yao, Liang
Yang, Jian
Tai, Ying
contents Subject-consistent generation (SCG)-aiming to maintain a consistent subject identity across diverse scenes-remains a challenge for text-to-image (T2I) models. Existing training-free SCG methods often achieve consistency at the cost of layout and pose diversity, hindering expressive visual storytelling. To address the limitation, we propose subject-Consistent and pose-Diverse T2I framework, dubbed as CoDi, that enables consistent subject generation with diverse pose and layout. Motivated by the progressive nature of diffusion, where coarse structures emerge early and fine details are refined later, CoDi adopts a two-stage strategy: Identity Transport (IT) and Identity Refinement (IR). IT operates in the early denoising steps, using optimal transport to transfer identity features to each target image in a pose-aware manner. This promotes subject consistency while preserving pose diversity. IR is applied in the later denoising steps, selecting the most salient identity features to further refine subject details. Extensive qualitative and quantitative results on subject consistency, pose diversity, and prompt fidelity demonstrate that CoDi achieves both better visual perception and stronger performance across all metrics. The code is provided in https://github.com/NJU-PCALab/CoDi.
format Preprint
id arxiv_https___arxiv_org_abs_2507_08396
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle CoDi: Subject-Consistent and Pose-Diverse Text-to-Image Generation
Gao, Zhanxin
Zhu, Beier
Yao, Liang
Yang, Jian
Tai, Ying
Computer Vision and Pattern Recognition
Subject-consistent generation (SCG)-aiming to maintain a consistent subject identity across diverse scenes-remains a challenge for text-to-image (T2I) models. Existing training-free SCG methods often achieve consistency at the cost of layout and pose diversity, hindering expressive visual storytelling. To address the limitation, we propose subject-Consistent and pose-Diverse T2I framework, dubbed as CoDi, that enables consistent subject generation with diverse pose and layout. Motivated by the progressive nature of diffusion, where coarse structures emerge early and fine details are refined later, CoDi adopts a two-stage strategy: Identity Transport (IT) and Identity Refinement (IR). IT operates in the early denoising steps, using optimal transport to transfer identity features to each target image in a pose-aware manner. This promotes subject consistency while preserving pose diversity. IR is applied in the later denoising steps, selecting the most salient identity features to further refine subject details. Extensive qualitative and quantitative results on subject consistency, pose diversity, and prompt fidelity demonstrate that CoDi achieves both better visual perception and stronger performance across all metrics. The code is provided in https://github.com/NJU-PCALab/CoDi.
title CoDi: Subject-Consistent and Pose-Diverse Text-to-Image Generation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2507.08396