One Framework to Register Them All: PointNet Encoding for Point Cloud Alignment

Abstract

PointNet has recently emerged as a popular representation for unstructured point cloud data, allowing application of deep learning to tasks such as object detection, segmentation and shape completion. However, recent works in literature have shown the sensitivity of the PointNet representation to pose misalignment. This paper presents a novel framework that uses PointNet encoding to align point clouds and perform registration for applications such as 3D reconstruction, tracking and pose estimation. We develop a framework that compares PointNet features of template and source point clouds to find the transformation that aligns them accurately. In doing so, we avoid computationally expensive correspondence finding steps, that are central to popular registration methods such as ICP and its variants. Depending on the prior information about the shape of the object formed by the point clouds, our framework can produce approaches that are shape specific or general to unseen shapes. Our framework produces approaches that are robust to noise and initial misalignment in data and work robustly with sparse as well as partial point clouds. We perform extensive simulation and real-world experiments to validate the efficacy of our approach and compare the performance with state-of-art approaches. Code is available at https://github.com/vinits5/pointnet-registrationframework.

10 pages. arXiv admin note: substantial text overlap with arXiv:1908.07906

References in corpus (6)

• Go-ICP: A Globally Optimal Solution to 3D ICP Point-Set Registration
• 3DFeat-Net: Weakly Supervised Local 3D Features for Point Cloud Registration
• Dynamic Graph CNN for Learning on Point Clouds
• Deep Closest Point: Learning Representations for Point Cloud Registration
• Iterative Transformer Network for 3D Point Cloud
• PCN: Point Completion Network

Cited by in corpus (4)

• VRNet: Learning the Rectified Virtual Corresponding Points for 3D Point Cloud Registration
• Pillar-based Object Detection for Autonomous Driving
• Robust Partial-to-Partial Point Cloud Registration in a Full Range
• A Representation Separation Perspective to Correspondences-free Unsupervised 3D Point Cloud Registration