PWCLO-Net: Deep LiDAR Odometry in 3D Point Clouds Using Hierarchical Embedding Mask Optimization

  title={PWCLO-Net: Deep LiDAR Odometry in 3D Point Clouds Using Hierarchical Embedding Mask Optimization},
  author={Guangming Wang and Xin Wu and Zhe Liu and Hesheng Wang},
  journal={2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  • Guangming WangXin Wu Hesheng Wang
  • Published 2 December 2020
  • Computer Science
  • 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
A novel 3D point cloud learning model for deep LiDAR odometry, named PWCLO-Net, using hierarchical embedding mask optimization is proposed in this paper. In this model, the Pyramid, Warping, and Cost volume (PWC) structure for the LiDAR odometry task is built to refine the estimated pose in a coarse-to-fine approach hierarchically. An attentive cost volume is built to associate two point clouds and obtain embedding motion patterns. Then, a novel trainable embedding mask is proposed to weigh the… 

Figures and Tables from this paper

Attention Models for Point Clouds in Deep Learning: A Survey

This paper provides a detailed characterization of the role of attention mechanisms, the usability of attention models into different tasks, and the development trend of key technology in 3D point clouds feature representation.

Pseudo-LiDAR for Visual Odometry

—As one of the important tasks in the field of robotics and machine vision, LiDAR/visual odometry provides tremendous help for various applications such as navigation, location, etc. In the existing

Efficient 3D Deep LiDAR Odometry

The proposed EfficientLO-Net architecture outperforms all recent learning-based methods and even the geometry-based approach, LOAM with mapping optimization, on most sequences of KITTI odometry dataset.

NeRF-LOAM: Neural Implicit Representation for Large-Scale Incremental LiDAR Odometry and Mapping

  • Junyuan DengXieyuanli Chen Ling Pei
  • Environmental Science
  • 2023
Simultaneously odometry and mapping using LiDAR data is an important task for mobile systems to achieve full autonomy in large-scale environments. However, most existing LiDAR-based methods

RegFormer: An Efficient Projection-Aware Transformer Network for Large-Scale Point Cloud Registration

  • Jiuming LiuGuangming WangZhe LiuChaokang JiangMarc PollefeysHesheng Wang
  • Computer Science
  • 2023
This work proposes an end-to-end transformer network (RegFormer) for large-scale point cloud alignment without any further post-processing, and proposes a projection-aware hierarchical transformer to capture long-range dependencies and filter outliers by extracting point features globally.

LoRCoN-LO: Long-term Recurrent Convolutional Network-based LiDAR Odometry

A deep learning-based LiDAR odometry estimation method called LoRCoN-LO that utilizes the long-term recurrent convolutional network (LRCN) structure that is suitable for predicting continuous robot movements as it uses point clouds that contain spatial information.

Deep Planar Parallax for Monocular Depth Estimation

Comprehensive experimental results on autonomous driving datasets like KITTI and Waymo Open Dataset (WOD) demonstrate that the Planar Parallax Network (PPNet) dramatically outperforms existing learning-based methods.

Pseudo-Anchors: Robust Semantic Features for Lidar Mapping in Highly Dynamic Scenarios

This study imitates anchor-based approaches such as magnetic nails by applying novel Static Confidence Criteria (SCC) to the point-cloud semantic candidates to ensure their robustness by name Pseudo-Anchors (P-A) as they hold similar properties to the anchor nodes.

STCLoc: Deep LiDAR Localization With Spatio-Temporal Constraints

A novel LiDAR localization framework with spatio-temporal constraints is proposed, termed STCLoc, to reduce scene ambiguities and achieve more accurate localization, and is proposed to regularize regression in the spatial dimension with a novel classification task to reduce outliers.

A 3D LiDAR odometry for UGVs using coarse-to-fine deep scene flow estimation

Light detection and ranging (LiDAR) odometry plays a crucial role in autonomous mobile robots and unmanned ground vehicles (UGVs). This paper presents a deep learning–based odometry system using two



LodoNet: A Deep Neural Network with 2D Keypoint Matching for 3D LiDAR Odometry Estimation

This work transfers the LiDAR frames to image space and reformulate the problem as image feature extraction with the help of scale-invariant feature transform (SIFT) for feature extraction, and is able to generate matched keypoint pairs (MKPs) that can be precisely returned to the 3D space.

PointPWC-Net: A Coarse-to-Fine Network for Supervised and Self-Supervised Scene Flow Estimation on 3D Point Clouds

This work proposes a novel end-to-end deep scene flow model, called PointPWC-Net, on 3D point clouds in a coarse- to-fine fashion, which shows great generalization ability on KITTI Scene Flow 2015 dataset, outperforming all previous methods.

DeepPCO: End-to-End Point Cloud Odometry through Deep Parallel Neural Network

This work investigates how to exploit deep learning to estimate point cloud odometry (PCO), which may serve as a critical component in point cloud-based downstream tasks or learning-based systems, and proposes a novel end-to-end deep parallel neural network called DeepPCO, which can estimate the 6-DOF poses using consecutive point clouds.

Hierarchical Attention Learning of Scene Flow in 3D Point Clouds

A novel hierarchical neural network with double attention is proposed for learning the correlation of point features in adjacent frames and refining scene flow from coarse to fine layer by layer, which outperforms the ICP-based method and shows good practical application ability.

FlowNet3D: Learning Scene Flow in 3D Point Clouds

  • Xingyu LiuC. QiL. Guibas
  • Computer Science
    2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
  • 2019
This work proposes a novel deep neural network named FlowNet3D that learns scene flow from point clouds in an end-to-end fashion and successfully generalizes to real scans, outperforming various baselines and showing competitive results to the prior art.

LO-Net: Deep Real-Time Lidar Odometry

A novel deep convolutional network pipeline, LO-Net, for real-time lidar odometry estimation that outperforms existing learning based approaches and has similar accuracy with the state-of-the-art geometry-based approach, LOAM.

Collar Line Segments for fast odometry estimation from Velodyne point clouds

Evaluation using the KITTI dataset shows that the method outperforms publicly available and commonly used state-of-the-art method GICP for point cloud registration in both accuracy and speed, especially in cases where the scene lacks significant landmarks or in typical urban elements.

Multi-view 3D Object Detection Network for Autonomous Driving

This paper proposes Multi-View 3D networks (MV3D), a sensory-fusion framework that takes both LIDAR point cloud and RGB images as input and predicts oriented 3D bounding boxes and designs a deep fusion scheme to combine region-wise features from multiple views and enable interactions between intermediate layers of different paths.

CNN for IMU assisted odometry estimation using velodyne LiDAR

This work introduces a novel method for odometry estimation using convolutional neural networks from 3D LiDAR scans and proposes alternative CNNs trained for the prediction of rotational motion parameters while achieving results also comparable with state of the art method LOAM.

Unsupervised Geometry-Aware Deep LiDAR Odometry

This work focuses on unsupervised learning for LiDAR odometry (LO) without trainable labels, and introduces the uncertainty-aware loss with geometric confidence, thereby al-lowing the reliability of the proposed pipeline.