# SelfVoxeLO: Self-supervised LiDAR Odometry with Voxel-based Deep Neural Networks

arXiv: https://arxiv.org/abs/2010.09343 · HTML: https://mlsanity.com/paper/2010.09343/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 19 Oct 2020.

## Built on by (most cited first, 10 of 10)

- [Improving Visual Grounding with Visual-Linguistic Verification and Iterative Reasoning](https://mlsanity.com/paper/2205.00272/) (2022)
- [RegFormer: An Efficient Projection-Aware Transformer Network for Large-Scale Point Cloud Registration](https://mlsanity.com/paper/2303.12384/) (2023)
- [VS-Net: Voting with Segmentation for Visual Localization](https://mlsanity.com/paper/2105.10886/) (2021)
- [Robust Self-Supervised LiDAR Odometry via Representative Structure Discovery and 3D Inherent Error Modeling](https://mlsanity.com/paper/2202.13353/) (2022)
- [UnDeepLIO: Unsupervised Deep Lidar-Inertial Odometry](https://mlsanity.com/paper/2109.01533/) (2021)
- [RegFormer++: An Efficient Large-Scale 3D LiDAR Point Registration Network with Projection-Aware 2D Transformer](https://mlsanity.com/paper/2603.14290/) (2026)
- [Slice Transformer and Self-supervised Learning for 6DoF Localization in 3D Point Cloud Maps](https://mlsanity.com/paper/2301.08957/) (2023)
- [UnPWC-SVDLO: Multi-SVD on PointPWC for Unsupervised Lidar Odometry](https://mlsanity.com/paper/2205.08150/) (2022)
- [Attention-Guided Lidar Segmentation and Odometry Using Image-to-Point Cloud Saliency Transfer](https://mlsanity.com/paper/2308.14332/) (2023)
- [Efficient Active Training for Deep LiDAR Odometry](https://mlsanity.com/paper/2509.03211/) (2025)

## Abstract (arXiv)

Recent learning-based LiDAR odometry methods have demonstrated their competitiveness. However, most methods still face two substantial challenges: 1) the 2D projection representation of LiDAR data cannot effectively encode 3D structures from the point clouds; 2) the needs for a large amount of labeled data for training limit the application scope of these methods. In this paper, we propose a self-supervised LiDAR odometry method, dubbed SelfVoxeLO, to tackle these two difficulties. Specifically, we propose a 3D convolution network to process the raw LiDAR data directly, which extracts features that better encode the 3D geometric patterns. To suit our network to self-supervised learning, we design several novel loss functions that utilize the inherent properties of LiDAR point clouds. Moreover, an uncertainty-aware mechanism is incorporated in the loss functions to alleviate the interference of moving objects/noises. We evaluate our method's performances on two large-scale datasets, i.e., KITTI and Apollo-SouthBay. Our method outperforms state-of-the-art unsupervised methods by 27%/32% in terms of translational/rotational errors on the KITTI dataset and also performs well on the Apollo-SouthBay dataset. By including more unlabelled training data, our method can further improve performance comparable to the supervised methods.