Deep Closest Point: Learning Representations for Point Cloud Registration

Abstract

Point cloud registration is a key problem for computer vision applied to robotics, medical imaging, and other applications. This problem involves finding a rigid transformation from one point cloud into another so that they align. Iterative Closest Point (ICP) and its variants provide simple and easily-implemented iterative methods for this task, but these algorithms can converge to spurious local optima. To address local optima and other difficulties in the ICP pipeline, we propose a learning-based method, titled Deep Closest Point (DCP), inspired by recent techniques in computer vision and natural language processing. Our model consists of three parts: a point cloud embedding network, an attention-based module combined with a pointer generation layer, to approximate combinatorial matching, and a differentiable singular value decomposition (SVD) layer to extract the final rigid transformation. We train our model end-to-end on the ModelNet40 dataset and show in several settings that it performs better than ICP, its variants (e.g., Go-ICP, FGR), and the recently-proposed learning-based method PointNetLK. Beyond providing a state-of-the-art registration technique, we evaluate the suitability of our learned features transferred to unseen objects. We also provide preliminary analysis of our learned model to help understand whether domain-specific and/or global features facilitate rigid registration.

References in corpus (10)

• Sequence to Sequence Learning with Neural Networks
• Semi-Supervised Classification with Graph Convolutional Networks
• Geometric deep learning: going beyond Euclidean data
• Open3D: A Modern Library for 3D Data Processing
• Go-ICP: A Globally Optimal Solution to 3D ICP Point-Set Registration
• Layer Normalization
• Dynamic Graph CNN for Learning on Point Clouds
• Relational Deep Reinforcement Learning
• Feature Denoising for Improving Adversarial Robustness
• Relation Networks for Object Detection

Cited by in corpus (16)

• A Smooth Representation of Belief over SO(3) for Deep Rotation Learning with Uncertainty
• Under the Radar: Learning to Predict Robust Keypoints for Odometry Estimation and Metric Localisation in Radar
• Iterative Distance-Aware Similarity Matrix Convolution with Mutual-Supervised Point Elimination for Efficient Point Cloud Registration
• One Framework to Register Them All: PointNet Encoding for Point Cloud Alignment
• Unsupervised Learning of 3D Point Set Registration
• A Learnable Self-supervised Task for Unsupervised Domain Adaptation on Point Clouds
• PosDiffNet: Positional Neural Diffusion for Point Cloud Registration in a Large Field of View with Perturbations
• DMLO: Deep Matching LiDAR Odometry
• Attention Models for Point Clouds in Deep Learning: A Survey
• Unsupervised Learning of Global Registration of Temporal Sequence of Point Clouds
• Sampling Network Guided Cross-Entropy Method for Unsupervised Point Cloud Registration
• Deep-3DAligner: Unsupervised 3D Point Set Registration Network With Optimizable Latent Vector
• Irregularly Tabulated MLP for Fast Point Feature Embedding
• GP-Aligner: Unsupervised Non-rigid Groupwise Point Set Registration Based On Optimized Group Latent Descriptor
• Review on 6D Object Pose Estimation with the focus on Indoor Scene Understanding
• Fast Simultaneous Gravitational Alignment of Multiple Point Sets