Redirecting to original paper in 30 seconds...
Click below to go immediately or wait for automatic redirect
📄 Abstract
Abstract: Despite their impressive realism, modern text-to-image models still struggle
with compositionality, often failing to render accurate object counts,
attributes, and spatial relations. To address this challenge, we present a
training-free framework that combines an object-centric approach with
self-refinement to improve layout faithfulness while preserving aesthetic
quality. Specifically, we leverage large language models (LLMs) to synthesize
explicit layouts from input prompts, and we inject these layouts into the image
generation process, where a object-centric vision-language model (VLM) judge
reranks multiple candidates to select the most prompt-aligned outcome
iteratively. By unifying explicit layout-grounding with self-refine-based
inference-time scaling, our framework achieves stronger scene alignment with
prompts compared to recent text-to-image models. The code are available at
https://github.com/gcl-inha/ReFocus.
Authors (3)
Minsuk Ji
Sanghyeok Lee
Namhyuk Ahn
Submitted
October 28, 2025
Key Contributions
This paper presents a training-free framework to improve compositionality in text-to-image synthesis. It uses LLMs to generate explicit layouts from prompts, injects these layouts into the generation process, and employs an object-centric VLM to rerank candidates, achieving stronger scene alignment with prompts through inference-time scaling and self-refinement.
Business Value
Enables the creation of more precise and controllable visual content from text descriptions, benefiting graphic designers, advertisers, and content creators who require accurate scene composition.