Year: 2025
Type: preprint
Abstract: Personalized image synthesis has emerged as a pivotal application in text-to-image generation, enabling the creation of images featuring specific subjects in diverse contexts. While diffusion models have dominated this domain, auto-regressive models, with their unified architecture for text and image modeling, remain underexplored for personalized image generation. This paper investigates the potential of optimizing auto-regressive models for personalized image synthesis, leveraging their inherent multimodal capabilities to perform this task. We propose a two-stage training strategy that combines optimization of text embeddings and fine-tuning of transformer layers. Our experiments on the auto-regressive model demonstrate that this method achieves comparable subject fidelity and prompt following to the leading diffusion-based personalization methods. The results highlight the effectiveness of auto-regressive models in personalized image generation, offering a new direction for future research in this area.
Source: arXiv (Cornell University)
Authors: Sun, Kaiyue, Liu, Xian, Teng, Yao, Liu, Xihui
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W4416538112
Cites:
Cited by:
Related to: