DOSE3: Diffusion-Based Unified Out-Of-Distribution Detection on SE(3) Trajectories
Hongzhe Cheng, Tianyou Zheng, Ziyong Ma, Tianyi Zhang, Matthew Johnson-Roberson, Weiming Zhi
AI summary
Problem
Existing unified out-of-distribution detection methods are limited to Euclidean spaces and require retraining for new data distributions, leaving a critical gap for safely evaluating 3D rigid-body motion trajectories in robotics.
Approach
The authors train a single unconditional diffusion model directly on the SE(3) manifold to process 6-DOF pose sequences, then extract a high-dimensional OOD metric from the model's noise estimator without retraining.
Key results
- A unified diffusion framework that respects SE(3) manifold geometry for pose sequences
- A novel 24-dimensional OOD statistic derived from diffusion noise estimators
- Strong OOD detection performance across autonomous driving and manipulator datasets
- Zero-shot adaptability to new inlier trajectory distributions without retraining
Why it matters
Provides a scalable, retraining-free safety mechanism for autonomous vehicles and robots operating in complex 3D environments.
Abstract
Out-of-Distribution (OOD) detection, the task of identifying when an input falls outside the distribution seen at training time, is critical for deploying safe and reliable systems. Traditional OOD methods require retraining models whenever the in-distribution has changed. Recent work introduces unified models for OOD detection, where metrics can be constructed from an unconditional diffusion model trained on an arbitrary dataset, and the inlier distribution can be changed without retraining the diffusion model. However, these unified approaches have been largely confined to Euclidean or latent space domains. In contrast, real-world robotics systems often perceive and act through sequences of 6 degrees-of-freedom poses in the Special Euclidean Group SE(3), taking into account both translations and orientation changes over time. In this work, we extend OOD detection to trajectories in Special Euclidean Group in 3D (SE(3)) by presenting a Diffusion-based Out-of-distribution detection on SE(3) (DOSE3). DOSE3 constructs an OOD metric from the noise estimator model of a diffusion model over SE(3) to separate outlier samples from inlier distributions. We demonstrate DOSE3’s strong performance on OOD detection frameworks through extensive validation on multiple real-world robotics and autonomous systems datasets, covering vehicle and robot manipulator motion trajectories.