Full-DoF Egomotion Estimation for Event Cameras Using Geometric Solvers

2025-03-05

AI Summary

This work addresses the limitation of existing sparse geometric solvers for event cameras—which rely on external rotational priors (e.g., IMU) and estimate only translation—by proposing, for the first time, a motion prior-free framework for full 6-DOF egomotion estimation. The core method introduces a line-segment-induced event manifold geometric model and a novel normal-vector coplanarity constraint, eliminating the conventional requirement of known rotation. By combining first-order rotational approximation with Adam optimization, angular and linear velocities are jointly optimized directly on the event manifold. Extensive evaluations on both synthetic and real-world event data demonstrate high pose estimation accuracy and robustness under challenging conditions. To our knowledge, this is the first framework achieving full 6-DOF motion estimation from event cameras using only sparse geometric constraints—without auxiliary sensors or motion priors. The source code is publicly available.

Technology Category

• Computer Vision: Motion & Tracking
• Intelligent Robots: State Estimation
• Machine Learning: Learning with Manifolds

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

For event cameras, current sparse geometric solvers for egomotion estimation assume that the rotational displacements are known, such as those provided by an IMU. Thus, they can only recover the translational motion parameters. Recovering full-DoF motion parameters using a sparse geometric solver is a more challenging task, and has not yet been investigated. In this paper, we propose several solvers to estimate both rotational and translational velocities within a unified framework. Our method leverages event manifolds induced by line segments. The problem formulations are based on either an incidence relation for lines or a novel coplanarity relation for normal vectors. We demonstrate the possibility of recovering full-DoF egomotion parameters for both angular and linear velocities without requiring extra sensor measurements or motion priors. To achieve efficient optimization, we exploit the Adam framework with a first-order approximation of rotations for quick initialization. Experiments on both synthetic and real-world data demonstrate the effectiveness of our method. The code is available at https://github.com/jizhaox/relpose-event.

Problem

• Estimates full-DoF egomotion for event cameras
• Recovers rotational and translational velocities
• Uses geometric solvers without extra sensors

Innovation

• Estimates full-DoF motion using event manifolds
• Employs geometric solvers for rotational and translational velocities
• Utilizes Adam framework for efficient optimization