驯服I2V模型用于图像HOI编辑：认知基准与智能体自校正框架
Taming I2V models for Image HOI Editing: A Cognitive Benchmark and Agentic Self-Correcting Framework
AI总结：
提出HOI-Edit基准和SCPE框架，利用I2V模型的时间生成能力进行动态人-物交互编辑，通过自校正提示迭代优化，实现与SOTA竞争的性能。
摘要：
当前的图像编辑方法在静态属性上表现出色，但在复杂的人-物交互（HOI）上失败，这是一个关键挑战，现有基准将HOI与静态属性混淆，依赖无法同时评估动态交互有效性和纠缠的人-物对保留的全局指标。因此，我们首先引入HOI-Edit，一个包含三个渐进认知层次的综合基准，其特点是自动化指标HOI-Eval，通过让VLM在思考后对包含基础人-物对的图像进行问答，可靠地评估实例级交互。考虑到任务本质是重塑动态关系，我们对图像到视频（I2V）模型进行基准测试，发现它们由于其时间生成能力而天生适合动态编辑。关键的是，除了优越的性能，这种能力提供了“失败过程的重放”，为错误原因提供了独特的可诊断性。因此，我们提出SCPE（自校正过程编辑），一种新颖的智能体自校正框架，通过迭代优化的提示约束I2V模型的生成，使生成的视频更准确地呈现目标HOI。从这些视频中提取的帧是最终的编辑结果。在HOI-Edit上，SCPE在交互上达到了与最先进（SOTA）编辑模型（如Nano Banana）竞争的性能。代码可在该https URL获取。
英文摘要：
Current image editing methods excel at static attributes but fail at complex Human-Object Interactions (HOI), a critical challenge unaddressed by existing benchmarks that conflate HOI with static attributes, relying on global metrics incapable of simultaneously assessing dynamic interaction validity and entangled human-object pair preservation. Thus, we first introduce HOI-Edit, a comprehensive benchmark with three progressive cognitive levels, which features an automated metric HOI-Eval that reliably evaluates instance-level interaction by letting VLM Q&A after thinking with images containing grounded Human-Object pairs. Considering the task's essence of remodeling dynamic relationships, we benchmark Image-to-Video (I2V) models, finding them inherently suited for dynamic editing due to their temporal generation capabilities. Crucially, beyond superior performance, this capability provides a "replay of the failure process," offering unique diagnosability into why errors occur. We thus propose SCPE (Self-Correcting Process Editing), a novel, agentic self-correcting framework that constrains the generation of I2V models through iteratively refined prompts, enabling the generated videos to more accurately present the target HOI. Extracted frames from these videos are the final editing results. On HOI-Edit, SCPE achieves performance competitive with state-of-the-art (SOTA) editing models like Nano Banana on interaction. Code is available at https://github.com/oceanflowlab/HOI-Edit.
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。
