PRNet: Self-Supervised Learning for Partial-to-Partial Registration

Abstract

We present a simple, flexible, and general framework titled Partial Registration Network (PRNet), for partial-to-partial point cloud registration. Inspired by recently-proposed learning-based methods for registration, we use deep networks to tackle non-convexity of the alignment and partial correspondence problems. While previous learning-based methods assume the entire shape is visible, PRNet is suitable for partial-to-partial registration, outperforming PointNetLK, DCP, and non-learning methods on synthetic data. PRNet is self-supervised, jointly learning an appropriate geometric representation, a keypoint detector that finds points in common between partial views, and keypoint-to-keypoint correspondences. We show PRNet predicts keypoints and correspondences consistently across views and objects. Furthermore, the learned representation is transferable to classification.

NeurIPS 2019

Cited by in corpus (32)

• RPM-Net: Robust Point Matching using Learned Features
• A comprehensive survey on point cloud registration
• DPFM: Deep Partial Functional Maps
• PRS-Net: Planar Reflective Symmetry Detection Net for 3D Models
• TEASER: Fast and Certifiable Point Cloud Registration
• Pillar-based Object Detection for Autonomous Driving
• Multi-Features Guidance Network for partial-to-partial point cloud registration
• You Only Hypothesize Once: Point Cloud Registration with Rotation-equivariant Descriptors
• DeepGMR: Learning Latent Gaussian Mixture Models for Registration
• PointDSC: Robust Point Cloud Registration using Deep Spatial Consistency
• Robust Point Cloud Registration Framework Based on Deep Graph Matching
• Robust Partial-to-Partial Point Cloud Registration in a Full Range
• Correspondence-Free Point Cloud Registration with SO(3)-Equivariant Implicit Shape Representations
• Correspondence Matrices are Underrated
• Learning 3D-3D Correspondences for One-shot Partial-to-partial Registration
• PREDATOR: Registration of 3D Point Clouds with Low Overlap
• Tabulated MLP for Fast Point Feature Embedding
• Correspondence Learning via Linearly-invariant Embedding
• Fast Gravitational Approach for Rigid Point Set Registration with Ordinary Differential Equations
• Mapping in a cycle: Sinkhorn regularized unsupervised learning for point cloud shapes
• Planning with Learned Dynamic Model for Unsupervised Point Cloud Registration
• Irregularly Tabulated MLP for Fast Point Feature Embedding
• A Method to Generate High Precision Mesh Model and RGB-D Datasetfor 6D Pose Estimation Task
• Registration Loss Learning for Deep Probabilistic Point Set Registration
• End-to-End 3D Point Cloud Learning for Registration Task Using Virtual Correspondences
• ASFlow: Unsupervised Optical Flow Learning with Adaptive Pyramid Sampling
• 3D Registration for Self-Occluded Objects in Context
• What Stops Learning-based 3D Registration from Working in the Real World?
• Consistent Two-Flow Network for Tele-Registration of Point Clouds
• A Robust Loss for Point Cloud Registration
• Unsupervised Dense Deformation Embedding Network for Template-Free Shape Correspondence
• PRANet: Point Cloud Registration with an Artificial Agent