[论文] VoxProg: 3D Spatial Reasoning through Voxel Programming

论文概要

研究领域: CV
作者: Zhichao Zhai, Menglei Zhao, Guikun Chen, Zihao Yan, Kanzhi Wu, Wenguan Wang, Jun Xiao
发布时间: 2026-10-21
arXiv: 2610.30127

中文摘要

语言模型解释任务和生成程序的能力为 3D 空间推理提供了有前景的基础。然而在 3D 场景中执行空间指令时，程序需要构建语言中隐而未发的几何引用。为此我们引入 VoxProg，通过体素编程来构建 3D 物体放置：语言模型将任务理解转化为在场景表面局部体素视图上的可执行程序。由于这些视图以规则网格索引表面几何，具有显式坐标和邻域关系，证据程序可以直接从场景网格中分离出任务相关的区域和边界，下游姿态程序则可以测量所得几何来计算目标 SE(3) 姿态。执行产生的视觉和数值反馈使模型能够检查中间选择和候选放置，迭代修正代码。为评估涵盖功能使用、几何拟合和上下文布局的开放式放置，我们引入了包含 71 项任务的 VoxProg-Bench 基准，以及配套的感知意图对齐和 3D 物理有效性评估套件。在 VoxProg-Bench 和外部基准上，VoxProg 始终优于基于关系、关键点和网格的基线方法。

原文摘要

Language models abilities to interpret tasks and generate programs offer a promising basis for 3D spatial reasoning. Yet following spatial instructions in 3D scenes requires programs to construct geometric references that language leaves implicit. In this work, we introduce VoxProg, which formulates 3D object placement through voxel programming: a language model translates task interpretations into executable programs over local voxel views of scene surfaces. Because these views index surface geometry on regular grids with explicit coordinates and neighborhood relations, an evidence program can isolate task-relevant regions and boundaries directly from scene meshes, and a downstream pose program can measure the resulting geometry to compute a target SE(3) pose. Visual and numerical feedbac...

自动采集于 2026-10-23

#论文 #arXiv #CV #小凯