发表机构
Shanghai Jiao Tong University; Huawei（上海交通大学; 华为）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
SCULPT是一种3D部件生成框架，通过减法组合方式生成部件，解决了现有方法的边界问题，在PartObjaverse上实现了最优几何性能，还能完成细粒度纹理部件分解。
AI 中文摘要
感知部件的3D生成旨在创建作为完整对象具有一致性、同时暴露结构部件以用于编辑、材质分配、动画制作和复用的数字资产。现有方法将该结构置于原生生成循环之外：基于分割的方法对已生成的形状进行划分，而加法方法从预定义布局、框或标记合成部件，再将其协调为整体。前者保留了生成的几何结构，但在确定部件边界前就固定了对象；后者虽明确了部件数量，但常导致共享边界出现间隙、穿透和材质不连续等问题。本文提出SCULPT框架，通过减法组合解决这些挑战。给定以结构化3D隐空间表示的完整对象，SCULPT迭代应用联合分割预测器，生成一个提取的部件及剩余对象。该预测器基于图像和当前3D状态执行耦合去噪过程，使提取的部件与更新后的剩余部分共同生成，而非在生成后再协调。联合分割预测器在其原生稀疏3D支撑的并集上处理两个输出，允许相邻支撑重叠，而非强制不相交的体素划分。当剩余支撑变为空或达到固定安全上限时，迭代终止，使生成部件的数量在该界限内适配每个对象。大量实验表明，SCULPT在PartObjaverse上实现了最先进的几何性能，同时在部件组装后保持了强大的完整对象重建能力。在四个数据集图像、一个文本生成图像输入和一个真实世界照片上的结果进一步显示，该方法在基准测试之外还能实现细粒度的纹理部件分解。
英文摘要
Part-aware 3D generation aims to create digital assets that are coherent as complete objects while exposing structural parts for editing, material assignment, animation, and reuse. Existing methods impose this structure outside the native generation loop: segmentation-based methods partition an already generated shape, while additive methods synthesize parts from predefined layouts, boxes, or tokens and then reconcile them into a whole. The former preserves the generated geometry but fixes the object before part boundaries are determined; the latter exposes part cardinality but often leaves shared boundaries vulnerable to gaps, interpenetrations, and material discontinuities. In this paper, we propose SCULPT, a framework that addresses these challenges through subtractive composition. Given a complete object represented in a structured 3D latent space, SCULPT iteratively applies a joint split predictor to generate one extracted part together with the remaining object. The predictor performs a coupled denoising process conditioned on both the image and the current 3D state, so the extracted part and updated remainder are generated together rather than reconciled after generation. The joint split predictor processes both outputs on the union of their native sparse 3D supports, allowing neighboring supports to overlap rather than imposing a disjoint voxel partition. The rollout ends when the remainder support becomes empty or reaches a fixed safety cap, allowing the number of generated parts to adapt to each object within that bound. Extensive experiments demonstrate state-of-the-art geometry on PartObjaverse while preserving strong complete-object reconstruction after part assembly. Results on four dataset images, one text-to-image-generated input, and one real-world photograph further show fine-grained textured part decomposition beyond the benchmark.
CommentsProject page: https://sculpt-part.github.io/ Code: https://github.com/sculpt-part/SCULPT
