arXiv 2602.090842026-02-21cs.CV

Agent Banana: 基于代理思维与工具的高质量图像编辑

Agent Banana: High-Fidelity Image Editing with Agentic Thinking and Tooling

TAMU（德克萨斯大学）；Brown University（布朗大学）；UW-Madison（威斯康星大学麦迪逊分校）；UCSD（加州大学圣地亚哥分校）；USC（南加州大学）；xAI（xAI公司）；Adobe Research（Adobe研究院）；Meta AI；Princeton University（普林斯顿大学）

Ruijie Ye, Jiayi Zhang, Zhuoxin Liu, Zihao Zhu, Siyuan Yang, Li Li, Tianfu Fu, Franck Dernoncourt, Yue Zhao, Jiacheng Zhu, Ryan Rossi, Wenhao Chai, Zhengzhong Tu

备注：: Project Website: agent-banana.github.io

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

Agent Banana通过分层代理规划-执行框架实现高质量图像编辑，解决多轮次编辑一致性与背景忠实度问题，同时保持指令遵循能力。

摘要：

我们研究了基于指令的图像编辑在专业工作流程中的应用，并识别出三个持续存在的挑战：(i) 编辑器经常过度编辑，修改内容超出用户的意图；(ii) 现有的模型大多为单轮次，而多轮次编辑可以改变物体的忠实度；(iii) 在约1000分辨率的评估与实际工作流程中经常操作超高清图像（例如4K）不匹配。我们提出了Agent Banana，一种分层的代理规划-执行框架，用于高质量、物体感知、反思性编辑。Agent Banana引入了两个关键机制：(1) 上下文折叠，将长交互历史压缩成结构化的记忆，以实现稳定的长周期控制；(2) 图像层分解，通过局部层编辑来保持非目标区域，同时实现原分辨率输出。为了支持严格的评估，我们构建了HDD-Bench，一个高清、对话式基准，具有可验证的逐步目标和原生4K图像（11.8百万像素），用于诊断长周期失败。在HDD-Bench上，Agent Banana在多轮次一致性和背景忠实度（例如IC 0.871，SSIM-OM 0.84，LPIPS-OM 0.12）方面表现最佳，同时在指令遵循方面保持竞争力，并在标准单轮次编辑基准上也取得了强劲的表现。我们希望这项工作能推动可靠、专业级的代理图像编辑及其在实际工作流程中的整合。

英文摘要：

We study instruction-based image editing under professional workflows and identify three persistent challenges: (i) editors often over-edit, modifying content beyond the user's intent; (ii) existing models are largely single-turn, while multi-turn edits can alter object faithfulness; and (iii) evaluation at around 1K resolution is misaligned with real workflows that often operate on ultra high-definition images (e.g., 4K). We propose Agent Banana, a hierarchical agentic planner-executor framework for high-fidelity, object-aware, deliberative editing. Agent Banana introduces two key mechanisms: (1) Context Folding, which compresses long interaction histories into structured memory for stable long-horizon control; and (2) Image Layer Decomposition, which performs localized layer-based edits to preserve non-target regions while enabling native-resolution outputs. To support rigorous evaluation, we build HDD-Bench, a high-definition, dialogue-based benchmark featuring verifiable stepwise targets and native 4K images (11.8M pixels) for diagnosing long-horizon failures. On HDD-Bench, Agent Banana achieves the best multi-turn consistency and background fidelity (e.g., IC 0.871, SSIM-OM 0.84, LPIPS-OM 0.12) while remaining competitive on instruction following, and also attains strong performance on standard single-turn editing benchmarks. We hope this work advances reliable, professional-grade agentic image editing and its integration into real workflows.

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。