A Smooth Representation of Belief over SO(3) for Deep Rotation Learning with Uncertainty

Abstract

Accurate rotation estimation is at the heart of robot perception tasks such as visual odometry and object pose estimation. Deep neural networks have provided a new way to perform these tasks, and the choice of rotation representation is an important part of network design. In this work, we present a novel symmetric matrix representation of the 3D rotation group, SO(3), with two important properties that make it particularly suitable for learned models: (1) it satisfies a smoothness property that improves convergence and generalization when regressing large rotation targets, and (2) it encodes a symmetric Bingham belief over the space of unit quaternions, permitting the training of uncertainty-aware models. We empirically validate the benefits of our formulation by training deep neural rotation regressors on two data modalities. First, we use synthetic point-cloud data to show that our representation leads to superior predictive accuracy over existing representations for arbitrary rotation targets. Second, we use image data collected onboard ground and aerial vehicles to demonstrate that our representation is amenable to an effective out-of-distribution (OOD) rejection technique that significantly improves the robustness of rotation estimates to unseen environmental effects and corrupted input images, without requiring the use of an explicit likelihood loss, stochastic sampling, or an auxiliary classifier. This capability is key for safety-critical applications where detecting novel inputs can prevent catastrophic failure of learned models.

arXiv commentIn Proceedings of Robotics: Science and Systems (RSS'20), Corvallis , Oregon, USA, Jul. 12-16, 2020

References in corpus (7)

• Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles
• Deep Exploration via Bootstrapped DQN
• Deep Closest Point: Learning Representations for Point Cloud Registration
• On the Continuity of Rotation Representations in Neural Networks
• DPC-Net: Deep Pose Correction for Visual Localization
• Towards neural networks that provably know when they don't know
• A Quaternion-based Certifiably Optimal Solution to the Wahba Problem with Outliers

Cited by in corpus (7)

• Active Learning of Discrete-Time Dynamics for Uncertainty-Aware Model Predictive Control
• Wide-Baseline Relative Camera Pose Estimation with Directional Learning
• Intraoperative 2D/3D Image Registration via Differentiable X-ray Rendering
• Uncertainty-Aware Self-Supervised Learning of Spatial Perception Tasks
• Pose Representations for Deep Skeletal Animation
• Cryo-forum: A framework for orientation recovery with uncertainty measure with the application in cryo-EM image analysis
• Stable Object Reorientation using Contact Plane Registration