ReAgent: Point Cloud Registration using Imitation and Reinforcement Learning

TL;DR

ReAgent introduces a reinforcement learning-based approach for point cloud registration that outperforms classical and existing learning methods in accuracy and efficiency, with successful application to object pose estimation.

Contribution

The paper presents a novel registration agent using imitation and reinforcement learning, improving accuracy and inference speed over prior methods.

Findings

Achieves state-of-the-art registration accuracy on ModelNet40 and ScanObjectNN.

Reduces inference time compared to related approaches.

Outperforms state-of-the-art pose refinement methods on LINEMOD.

Abstract

Point cloud registration is a common step in many 3D computer vision tasks such as object pose estimation, where a 3D model is aligned to an observation. Classical registration methods generalize well to novel domains but fail when given a noisy observation or a bad initialization. Learning-based methods, in contrast, are more robust but lack in generalization capacity. We propose to consider iterative point cloud registration as a reinforcement learning task and, to this end, present a novel registration agent (ReAgent). We employ imitation learning to initialize its discrete registration policy based on a steady expert policy. Integration with policy optimization, based on our proposed alignment reward, further improves the agent's registration performance. We compare our approach to classical and learning-based registration methods on both ModelNet40 (synthetic) and ScanObjectNN (real data) and show that our ReAgent achieves state-of-the-art accuracy. The lightweight architecture of the agent, moreover, enables reduced inference time as compared to related approaches. In addition, we apply our method to the object pose estimation task on real data (LINEMOD), outperforming state-of-the-art pose refinement approaches.