论文/arXiv 2609.36940·29 天前

DispFlow-GS：面向单目可变形 3D Gaussian Splatting 的运动解耦位移流监督

针对单目可变形 3DGS 运动监督域差，提出位移流与运动解耦框架，在动态场景基准验证运动定位和一致性提升。

AI 整理

• 感知与导航
• 评测与安全

原文链接：DispFlow-GS: Displacement Flow Supervision with Motion Disentangling for Monocular Deformable 3D Gaussian Splatting（arXiv 2609.36940 · 发布日期：2026-09-29）

DispFlow-GS: Displacement Flow Supervision with Motion Disentangling for Monocular Deformable 3D Gaussian Splatting

针对单目可变形 3DGS 运动监督域差，提出位移流与运动解耦框架，在动态场景基准验证运动定位和一致性提升。

核心要点

• 现有渲染 Gaussian flow 与光流对齐的监督，在运动建模上改进有限，存在域差。
• 方法将每个 Gaussian 的 3D 位移投溅到图像平面，并用中间视图渲染解耦场景与相机运动。
• 提出 DRC 运动感知指标，衡量预测形变与渲染改进之间的对齐关系。
• 动态场景基准上，运动定位和运动-渲染一致性最高提升 39% 和 6%，图像指标约变 0.1%。

验证范围与局限：验证范围为动态场景基准上的实验，摘要未说明真机或仿真设置。

英文摘要 (Abstract)

Accurate dynamic scene reconstruction is important for robotic perception, where temporally consistent representations of dynamic environments are essential. Deformable 3D Gaussian Splatting (3DGS) models dynamic scenes through deformation fields, and recent methods incorporate motion supervision by aligning rendered Gaussian flow with optical flow. However, we find that such Gaussian-flow-based supervision provides only limited improvements in motion modeling. We identify a fundamental limitation of this supervision paradigm, namely a domain gap between rendered Gaussian flow and optical flow. To address this limitation, we propose a motion supervision framework built on Displacement Flow, which splats per-Gaussian 3D displacements onto the image plane to provide direct and stable optimization signals. We further disentangle scene motion from camera motion via intermediate-view rendering, enabling more reliable motion priors and targeted constraints on deformation and geometry. We also observe a discrepancy between motion fidelity and image-based evaluation, where improved motion awareness does not necessarily translate into better rendered image quality or higher image-based metric scores. Motivated by this mismatch, we introduce Deformation-Rendering Consistency (DRC), a motion-aware metric that measures the alignment between predicted deformation and rendering improvement. Experiments on dynamic scene benchmarks show substantial improvements in motion localization and motion--rendering consistency, reaching up to 39% and 6%, respectively, while image-based metrics change by only about 0.1%. These results confirm the observed mismatch between motion fidelity and image-based evaluation, demonstrating the significance of DRC for motion-aware evaluation.

作者：Thai Duy Nguyen, Haitian Zhang, Addison Lin Wang

来源

本页事实对应的公开页面，以原文为准。

1. arXiv 2609.36940：DispFlow-GS: Displacement Flow Supervision with Motion Disentangling for Monocular Deformable 3D Gaussian Splatting访问 2026-10-21