代理流引导与并行展开搜索用于空间感知的文本到图像生成
Agentic Flow Steering and Parallel Rollout Search for Spatially Grounded Text-to-Image Generation
AI总结：
本文提出AFS-Search框架，通过闭合环路机制和空间感知引导提升文本到图像生成的精度与效率，实现性能和速度的双重优化。
摘要：
精确的文本到图像（T2I）生成已取得显著成功，但受限于静态文本编码器的有限关系推理能力和开环采样中的误差累积。在没有实时反馈的情况下，普通微分方程轨迹中的初始语义歧义不可避免地演变为空间约束的随机偏差。为此，我们引入AFS-Search（代理流引导与并行展开搜索），一种基于FLUX.1-dev的免训练闭合环路框架。AFS-Search结合了免训练闭合环路并行展开搜索和流引导机制，利用视觉-语言模型（VLM）作为语义批评者，诊断中间潜在特征并动态引导速度场。此外，我们将T2I生成视为一个顺序决策过程，通过前瞻模拟探索多条轨迹，并基于VLM引导的奖励选择最优路径。进一步，我们提供了AFS-Search-Pro以实现更高性能和AFS-Search-Fast以实现更快的生成速度。实验结果表明，我们的AFS-Search-Pro显著提升了原始FLUX.1-dev的性能，在三个不同的基准上均取得最先进的结果。同时，AFS-Search-Fast也显著提升了性能，同时保持快速的生成速度。
英文摘要：
Precise Text-to-Image (T2I) generation has achieved great success but is hindered by the limited relational reasoning of static text encoders and the error accumulation in open-loop sampling. Without real-time feedback, initial semantic ambiguities during the Ordinary Differential Equation trajectory inevitably escalate into stochastic deviations from spatial constraints. To bridge this gap, we introduce AFS-Search (Agentic Flow Steering and Parallel Rollout Search), a training-free closed-loop framework built upon FLUX.1-dev. AFS-Search incorporates a training-free closed-loop parallel rollout search and flow steering mechanism, which leverages a Vision-Language Model (VLM) as a semantic critic to diagnose intermediate latents and dynamically steer the velocity field via precise spatial grounding. Complementarily, we formulate T2I generation as a sequential decision-making process, exploring multiple trajectories through lookahead simulations and selecting the optimal path based on VLM-guided rewards. Further, we provide AFS-Search-Pro for higher performance and AFS-Search-Fast for quicker generation. Experimental results show that our AFS-Search-Pro greatly boosts the performance of the original FLUX.1-dev, achieving state-of-the-art results across three different benchmarks. Meanwhile, AFS-Search-Fast also significantly enhances performance while maintaining fast generation speed.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
