SOE-Net: A Self-Attention and Orientation Encoding Network for Point Cloud based Place Recognition

Abstract

We tackle the problem of place recognition from point cloud data and introduce a self-attention and orientation encoding network (SOE-Net) that fully explores the relationship between points and incorporates long-range context into point-wise local descriptors. Local information of each point from eight orientations is captured in a PointOE module, whereas long-range feature dependencies among local descriptors are captured with a self-attention unit. Moreover, we propose a novel loss function called Hard Positive Hard Negative quadruplet loss (HPHN quadruplet), that achieves better performance than the commonly used metric learning loss. Experiments on various benchmark datasets demonstrate superior performance of the proposed network over the current state-of-the-art approaches. Our code is released publicly at https://github.com/Yan-Xia/SOE-Net.

arXiv commentAccepted by CVPR2021 (Oral)

References in corpus (5)

• Self-Attention Generative Adversarial Networks
• 3DFeat-Net: Weakly Supervised Local 3D Features for Point Cloud Registration
• PointSIFT: A SIFT-like Network Module for 3D Point Cloud Semantic Segmentation
• 3DMatch: Learning Local Geometric Descriptors from RGB-D Reconstructions
• DH3D: Deep Hierarchical 3D Descriptors for Robust Large-Scale 6DoF Relocalization

Cited by in corpus (2)

• TransLoc3D : Point Cloud based Large-scale Place Recognition using Adaptive Receptive Fields
• Attention Models for Point Clouds in Deep Learning: A Survey