XVerse: Consistent Multi-Subject Control of Identity and Semantic Attributes via DiT Modulation

Research paper by Bowen Chen and collaborators. Published 2025-06-26. Topics: Diffusion Transformers, DiTs, text-to-image generation, multi-subject controlled generation, reference images.

By Bowen Chen, Mengyi Zhao, Haomiao Sun, Li Chen, Xu Wang, Kang Du, Xinglong WuResearch paper

Resource information

Authors: Bowen Chen, Mengyi Zhao, Haomiao Sun, Li Chen, Xu Wang, Kang Du, Xinglong Wu
Research area: Computer vision
Application domain: General AI research
AI methods: Transformers
Research topics: cs.CV
arXiv record date: 2025-06-26
Publication date: 2025-06-26
Publication: arXiv
Upstream last modified: Jun 27, 2025
Last record update: Nov 3, 2026

Abstract

Achieving fine-grained control over subject identity and semantic attributes (pose, style, lighting) in text-to-image generation, particularly for multiple subjects, often undermines the editability and coherence of Diffusion Transformers (DiTs). Many approaches introduce artifacts or suffer from attribute entanglement. To overcome these challenges, we propose a novel multi-subject controlled generation model XVerse. By transforming reference images into offsets for token-specific text-stream modulation, XVerse allows for precise and independent control for specific subject without disrupting image latents or features. Consequently, XVerse offers high-fidelity, editable multi-subject image synthesis with robust control over individual subject characteristics and semantic attributes. This advancement significantly improves personalized and complex scene generation capabilities.