EditHero：面向长时程部件级3D编辑与风格建模的基准
EditHero: A Benchmark for Long-Horizon Part-Level 3D Editing and Vibe Modeling
浏览论文内容
中文总结 AI 辅助
EditHero是首个长时程部件级3D编辑基准，通过确定性组装引擎和人工审查提供精确目标，对比非智能体与LLM/VLM智能体方法，发现后者更遵循指令且保留未编辑区域，但耗时较长。
中文摘要 AI 辅助
3D编辑方法通常在单一编辑任务上进行测试，然而一个资产是通过一系列长序列的修订构建而成的，每次修订都必须在实现所需更改的同时保持其他部分不变。我们提出了EditHero，据我们所知，这是首个面向长时程、部件级3D编辑的基准，包含自然语言指令以及几何和纹理的目标图像。一个确定性的组装引擎在每次编辑后生成精确的目标，并且每个序列都经过人工审查。我们使用EditHero来比较两种对立的3D编辑方法。非智能体方法自上而下地操作，从学习到的3D表示中重新生成对象，并推断需要保留的内容。相比之下，LLM/VLM智能体自下而上地操作，通过检查网格的代码进行编辑，仅重写指令所要求的部件。非智能体方法常常遗漏所需的更改，并干扰本应保持固定的区域。大多数LLM更严格地遵循指令，且所有LLM都能更好地保留未编辑的部分，但它们的每次编辑都需要数分钟。我们将发布该引擎和编辑序列，以支持对可靠迭代式3D编辑的研究。
英文摘要
3D editing methods are usually tested on a single edit, yet an asset is built through a long sequence of revisions, each of which must implement the requested change while leaving everything else unchanged. We introduce EditHero, to our knowledge the first benchmark for long-horizon, part-level 3D editing, with natural-language instructions and target images for both geometry and texture. A deterministic assembly engine produces the exact target after every edit, and every sequence is reviewed by hand. We use EditHero to compare 2 opposite approaches to 3D editing. Non-agentic methods operate top down, regenerating the object from a learned 3D representation and inferring what to keep. In contrast, LLM/VLM agents operate bottom up, editing through code that inspects the mesh and rewrites only the parts required by instructions. The non-agentic methods often miss the requested change and disturb regions that should stay fixed. Most LLMs follow instructions more closely, and all of them preserve the unedited parts better, but each of their edits takes minutes. We will release the engine and the edit sequences to support research on reliable iterative 3D editing.
发表机构
- Alaya Lab（Alaya实验室）
- The University of Tokyo（东京大学）
- Institute of Science Tokyo（东京科学大学）
- University of California, Merced（加州大学默塞德分校）
机构由 AI 辅助整理，请以论文原文为准。
