[2510.24133] Compositional Image Synthesis with Inference-Time Scaling

Computer Science > Computer Vision and Pattern Recognition arXiv:2510.24133 (cs) [Submitted on 28 Oct 2025 (v1), last revised 27 Mar 2026 (this version, v2)] Title:Compositional Image Synthesis with Inference-Time Scaling Authors:Minsuk Ji, Sanghyeok Lee, Namhyuk Ahn View a PDF of the paper titled Compositional Image Synthesis with Inference-Time Scaling, by Minsuk Ji and 2 other authors View PDF HTML (experimental) 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 this https URL. Comments: Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) Cite as: arXiv:2510.24133 [cs.CV]   (or arXiv:2510.24...