DisFlow: Scene Flow from Distance Field for Object Pose, Velocity Tracking, and Dynamic Object Reconstruction

2026-06-01

AI Summary

This work addresses the challenging problem of simultaneously estimating six-degree-of-freedom object poses, tracking motion, and reconstructing surfaces in dynamic scenes. The authors propose a method based on Gaussian Process Implicit Surfaces (GPIS) that models geometry using signed distance fields constrained by their normal derivatives, and estimates scene flow through inter-frame distance field variations. Building upon this formulation, they introduce a closed-form optimization for incremental point cloud registration and, for the first time, perform probabilistic fusion directly in the object’s coordinate frame. This tightly couples spatial and temporal information to preserve geometric consistency. The approach jointly outputs high-precision pose trajectories, velocities, dense surfaces, surface normals, and associated uncertainties, achieving both high-quality reconstruction and real-time performance on dynamic object sequences.

Abstract

We present \emph{DisFlow}, a novel framework for online scene flow estimation from distance field that enables \emph{6DoF dynamic object pose estimation}, \emph{motion tracking}, and \emph{surface reconstruction}. The scene is represented by Gaussian Process Implicit Surfaces (GPIS), with surface normals serving as derivative constraints, enabling accurate signed distance computations near the surface and gradient queries with uncertainty. With this representation as a foundation, we compute a scene flow from the distance field that describes how surface points are transported over time in consecutive frames. Through our flow, we can estimate an object's pose and motion by incrementally registering a new observed point cloud via an elegant closed-form optimisation. Unlike prior methods that operate in the camera or world frame, our approach performs probabilistic fusion directly in the \emph{object frame}, where the object remains geometrically consistent over time. The tight coupling of the DisFlow method in space and time yields dense geometry, surface normals, object pose trajectories, velocities, and uncertainty, all at real-time rates. We evaluate DisFlow on dynamic object sequences and demonstrate that it achieves accurate pose and motion tracking while simultaneously reconstructing high-quality object surfaces. Code publicly available at \href{https://github.com/LanWu076/disflow_ros2}{https://github.com/LanWu076/disflow\_ros2}

Problem

• scene flow
• distance field
• object pose estimation
• motion tracking
• dynamic object reconstruction

Innovation

• scene flow
• distance field
• Gaussian Process Implicit Surfaces
• 6DoF pose estimation
• dynamic object reconstruction