Procedura：智能體驅動的3D建模與程序化控制

摘要

原生3D生成器現在能從單張圖像還原出令人驚豔的網格幾何。然而，在機械加工物體應呈現銳利邊緣之處，稠密網格仍舊是柔軟的；它不具備零件分解結構，也未暴露任何用戶可編輯的參數。為了解決這些問題，我們探索「將3D形狀視為程式碼」的範式，利用並擴展大型語言模型在3D建模上的程式編寫能力。我們提出 Procedura，一個新穎的3D建模代理框架，能將物體撰寫為程序化組裝——一個參數化程式，其具名零件通過具有型別且可機器檢查的配合關係相互連接。根據文字提示，該代理將物體規劃為組裝圖，並逐零件編寫程式；每個放置位置都從配合的參考座標系求解，而非依賴猜測，且僅在編譯、配合與連通性檢查全部通過後才接受該零件。接著，一個解耦的視覺評論者會一次一項診斷修復，逐步精煉組裝。此外，同一張圖還承載每個零件的材質以及經模擬器驗證的關節結構。我們在 P3D-Bench 上以其組裝評判標準進行評估，並在我們自製的硬表面基準 MechBench-36 上使用同一評判標準。在兩項基準中，Procedura 在評判品質上均優於最先進的原生3D生成器及所有先前的3D程式碼代理；在我們評估的所有方法中，它產生了最銳利的邊緣，且是唯一輸出為可編輯、具零件結構程式的方案。

English

Native 3D generators now recover impressive mesh geometry from a single image. However, a dense mesh stays soft where a machined object should be sharp, it carries no part decomposition, and it exposes no parameter a user could edit. To address this, we explore the paradigm of 3D shape as code, leveraging and scaling the coding ability of an LLM for 3D modeling. We introduce Procedura, a novel 3D modeling agent framework that writes an object as a procedural assembly, a parametric program whose named parts are joined by typed, machine-checkable mates. From a text prompt, the agent plans the object as an assembly graph and writes the program part by part, solving each placement from the mated frames rather than guessing it, and admitting a part only once compile, mate, and connectivity checks pass. A decoupled vision critic then refines the assembly one diagnosed fix at a time. Moreover, the same graph carries per-part materials and a simulator-validated articulation. We evaluate on P3D-Bench under its assembly judge, and with the same judge on MechBench-36, our hard-surface benchmark. On both, Procedura outperforms state-of-the-art native 3D generators and every prior 3D-code agent on judged quality, produces the sharpest edges of any method we evaluate, and is the only one whose output is an editable, part-structured program.