LiDAR-UDA: Self-ensembling Through Time for Unsupervised LiDAR Domain
Adaptation
- URL: http://arxiv.org/abs/2309.13523v1
- Date: Sun, 24 Sep 2023 02:02:00 GMT
- Title: LiDAR-UDA: Self-ensembling Through Time for Unsupervised LiDAR Domain
Adaptation
- Authors: Amirreza Shaban, JoonHo Lee, Sanghun Jung, Xiangyun Meng, Byron Boots
- Abstract summary: We introduce LiDAR-UDA, a novel two-stage self-training-based Unsupervised Domain Adaptation (UDA) method for LiDAR segmentation.
We propose two techniques to reduce sensor discrepancy and improve pseudo label quality.
We evaluate our method on several public LiDAR datasets and show that it outperforms the state-of-the-art methods by more than $3.9%$ mIoU on average.
- Score: 22.206488779765234
- License: http://creativecommons.org/licenses/by/4.0/
- Abstract: We introduce LiDAR-UDA, a novel two-stage self-training-based Unsupervised
Domain Adaptation (UDA) method for LiDAR segmentation. Existing self-training
methods use a model trained on labeled source data to generate pseudo labels
for target data and refine the predictions via fine-tuning the network on the
pseudo labels. These methods suffer from domain shifts caused by different
LiDAR sensor configurations in the source and target domains. We propose two
techniques to reduce sensor discrepancy and improve pseudo label quality: 1)
LiDAR beam subsampling, which simulates different LiDAR scanning patterns by
randomly dropping beams; 2) cross-frame ensembling, which exploits temporal
consistency of consecutive frames to generate more reliable pseudo labels. Our
method is simple, generalizable, and does not incur any extra inference cost.
We evaluate our method on several public LiDAR datasets and show that it
outperforms the state-of-the-art methods by more than $3.9\%$ mIoU on average
for all scenarios. Code will be available at
https://github.com/JHLee0513/LiDARUDA.
Related papers
- LiDAR View Synthesis for Robust Vehicle Navigation Without Expert Labels [50.40632021583213]
We propose synthesizing additional LiDAR point clouds from novel viewpoints without physically driving at dangerous positions.
We train a deep learning model, which takes a LiDAR scan as input and predicts the future trajectory as output.
A waypoint controller is then applied to this predicted trajectory to determine the throttle and steering labels of the ego-vehicle.
arXiv Detail & Related papers (2023-08-02T20:46:43Z) - Refined Pseudo labeling for Source-free Domain Adaptive Object Detection [9.705172026751294]
Source-freeD is proposed to adapt source-trained detectors to target domains with only unlabeled target data.
Existing source-freeD methods typically utilize pseudo labeling, where the performance heavily relies on the selection of confidence threshold.
We present a category-aware adaptive threshold estimation module, which adaptively provides the appropriate threshold for each category.
arXiv Detail & Related papers (2023-03-07T08:31:42Z) - SeqOT: A Spatial-Temporal Transformer Network for Place Recognition
Using Sequential LiDAR Data [9.32516766412743]
We propose a transformer-based network named SeqOT to exploit the temporal and spatial information provided by sequential range images.
We evaluate our approach on four datasets collected with different types of LiDAR sensors in different environments.
Our method operates online faster than the frame rate of the sensor.
arXiv Detail & Related papers (2022-09-16T14:08:11Z) - BEVFusion: A Simple and Robust LiDAR-Camera Fusion Framework [20.842800465250775]
Current methods rely on point clouds from the LiDAR sensor as queries to leverage the feature from the image space.
We propose a surprisingly simple yet novel fusion framework, dubbed BEVFusion, whose camera stream does not depend on the input of LiDAR data.
We empirically show that our framework surpasses the state-of-the-art methods under the normal training settings.
arXiv Detail & Related papers (2022-05-27T06:58:30Z) - LiDAR Distillation: Bridging the Beam-Induced Domain Gap for 3D Object
Detection [96.63947479020631]
In many real-world applications, the LiDAR points used by mass-produced robots and vehicles usually have fewer beams than that in large-scale public datasets.
We propose the LiDAR Distillation to bridge the domain gap induced by different LiDAR beams for 3D object detection.
arXiv Detail & Related papers (2022-03-28T17:59:02Z) - Learning Moving-Object Tracking with FMCW LiDAR [53.05551269151209]
We propose a learning-based moving-object tracking method utilizing our newly developed LiDAR sensor, Frequency Modulated Continuous Wave (FMCW) LiDAR.
Given the labels, we propose a contrastive learning framework, which pulls together the features from the same instance in embedding space and pushes apart the features from different instances to improve the tracking quality.
arXiv Detail & Related papers (2022-03-02T09:11:36Z) - End-To-End Optimization of LiDAR Beam Configuration for 3D Object
Detection and Localization [87.56144220508587]
We take a new route to learn to optimize the LiDAR beam configuration for a given application.
We propose a reinforcement learning-based learning-to-optimize framework to automatically optimize the beam configuration.
Our method is especially useful when a low-resolution (low-cost) LiDAR is needed.
arXiv Detail & Related papers (2022-01-11T09:46:31Z) - Self-Supervised Person Detection in 2D Range Data using a Calibrated
Camera [83.31666463259849]
We propose a method to automatically generate training labels (called pseudo-labels) for 2D LiDAR-based person detectors.
We show that self-supervised detectors, trained or fine-tuned with pseudo-labels, outperform detectors trained using manual annotations.
Our method is an effective way to improve person detectors during deployment without any additional labeling effort.
arXiv Detail & Related papers (2020-12-16T12:10:04Z) - SelfVoxeLO: Self-supervised LiDAR Odometry with Voxel-based Deep Neural
Networks [81.64530401885476]
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.
We evaluate our method's performances on two large-scale datasets, i.e., KITTI and Apollo-SouthBay.
arXiv Detail & Related papers (2020-10-19T09:23:39Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.