MoRE: Motion-Regularized Encoding for Articulated 3D Generation

查看原文 / View Original 编辑 / Edit

Computer Science > Computer Vision and Pattern Recognition arXiv:2610.30337 (cs) [Submitted on 21 Oct 2026] Title: MoRE: Motion-Regularized Encoding for Articulated 3D Generation Authors: Guantian Zheng, Shan Jin.

计算机科学 > 计算机视觉与模式识别 arXiv:2610.30337 (cs) [提交于 2026 年 10 月 21 日] 标题：MoRE：用于关节式 3D 生成的运动正则化编码作者：Guantian Zheng, Shan Jin。

Abstract: Generating an articulated object from a single image requires recovering both its appearance and the mechanism governing its parts. A compact encoding of joint attributes does not by itself ensure that reconstruction preserves the resulting motion. We present MoRE, Motion-Regularized Encoding for articulated 3D generation.

摘要：从单张图像生成关节式物体需要同时恢复其外观和控制其部件的机制。关节属性的紧凑编码本身并不能确保重建过程保留由此产生的运动。我们提出了 MoRE，即用于关节式 3D 生成的运动正则化编码。

MoRE represents coordinate-aligned joint lines with axis classification and two axis-orthogonal Plücker moment coefficients, removing the arbitrary choice of an origin along the axis. During representation learning, decoded joint hypotheses are used to move corresponding three-dimensional anchors, and their motions are compared with those of annotated joints.

MoRE 通过轴分类和两个轴正交的普吕克（Plücker）矩系数来表示坐标对齐的关节线，消除了沿轴选择原点的任意性。在表征学习过程中，解码后的关节假设被用于移动相应的三维锚点，并将其运动与标注关节的运动进行比较。

A differentiable expectation over joint types and axis directions couples attribute reconstruction to the motion it induces. An image-conditioned flow learns the resulting functional latents, while a multi-state-adapted Gaussian model supplies appearance.

通过对关节类型和轴方向进行可微期望计算，将属性重建与其诱导的运动耦合起来。图像条件流学习由此产生的功能性潜在变量，而多状态自适应高斯模型则提供外观信息。

At inference, recovered part assignments and joints articulate a single set of rest-pose Gaussians without regenerating the object for each state. Experiments on PartNet-Mobility and ACD show stronger part-box agreement, articulated-state geometry and visual similarity than the compared baselines, while joint prediction reduces axis, origin and range errors relative to ArtiLatent.

在推理阶段，恢复的部件分配和关节使一组静止姿态的高斯模型产生关节运动，而无需为每个状态重新生成物体。在 PartNet-Mobility 和 ACD 上的实验表明，与对比基线相比，该方法在部件框一致性、关节状态几何结构和视觉相似度方面表现更强，同时关节预测在轴、原点和范围误差方面较 ArtiLatent 有所降低。

Controlled ablations support the joint-line encoding and show that motion supervision reduces anchor-motion error. Code and pretrained checkpoints will be released in the future.

受控消融实验支持了关节线编码的有效性，并表明运动监督减少了锚点运动误差。代码和预训练检查点将在未来发布。