A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
Filters
Deep Camera Pose Regression Using Pseudo-LiDAR
[article]
2022
arXiv
pre-print
Preserved Fulltext
We then propose FusionLoc, a novel architecture that uses pseudo-LiDAR to regress a 6DOF camera pose. ...
The results from this architecture are compared against various other state-of-the-art deep pose regression implementations using the 7 Scenes dataset. ...
Thus, our goals in this paper are as follows: 1) to prove the viability of pseudo-LiDAR within this space; and 2) to propose a novel architecture that uses pseudo-LiDAR to regress a 6DOF camera pose. ...
arXiv:2203.00080v1
fatcat:w6owofg27ffarosmlbqrsmumna
Multi-modal 3D Human Pose Estimation with 2D Weak Supervision in Autonomous Driving
[article]
2021
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
3D human pose estimation (HPE) in autonomous vehicles (AV) differs from other use cases in many factors, including the 3D resolution and range of data, absence of dense depth maps, failure modes for LiDAR ...
, relative location between the camera and LiDAR, and a high bar for estimation accuracy. ...
Point Network: The training loss for the regression
branch is a Huber loss Lreg on the generated pseudo 3D la-
bels, weighted by the reliability rk . ...
arXiv:2112.12141v1
fatcat:qxf3mbny3vdlffk4gn2awbwspi
Deep Learning on Monocular Object Pose Detection and Tracking: A Comprehensive Overview
[article]
2022
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
Other Versions
Among methods for object pose detection and tracking, deep learning is the most promising one that has shown better performance than others. ...
Therefore, this study presents a comprehensive review of recent progress in object pose detection and tracking that belongs to the deep learning technical route. ...
Although PL-MONO (Pseudo-LiDAR) has greatly promoted performance,a substantial performance gap exists between using pseudo-LiDAR and using real LiDAR. ...
arXiv:2105.14291v2
fatcat:2kxd4owthvf7tbcbnlqlqu4r3m
Self-Supervised Person Detection in 2D Range Data using a Calibrated Camera
[article]
2021
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
Other Versions
Faster R-CNN) on a calibrated camera to automatically generate training labels (called pseudo-labels) for 2D LiDAR-based person detectors. ...
However, only a few annotated datasets are available for training and testing these deep networks, potentially limiting their performance when deployed in new environments or with different LiDAR models ...
GENERATING PSEUDO-LABELS We use a calibrated camera to generate pseudo-labels for training a 2D LiDAR-based person detector. ...
arXiv:2012.08890v2
fatcat:mtkmjcc22jfn7c6a6vdfa55tgy
Autonomous Driving with Deep Learning: A Survey of State-of-Art Technologies
[article]
2020
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Other Versions
Almost at the same time, deep learning has made breakthrough by several pioneers, three of them (also called fathers of deep learning), Hinton, Bengio and LeCun, won ACM Turin Award in 2019. ...
This is a survey of autonomous driving technologies with deep learning methods. ...
[225] extend Pseudo-LiDAR with stereo camera and extremely sparse LiDAR sensor, called Pseudo-LiDAR++, in which a graph-based depth correction algorithm (GDC) refines the depth map by leveraging sparser ...
arXiv:2006.06091v3
fatcat:nhdgivmtrzcarp463xzqvnxlwq
DeepTIO: A Deep Thermal-Inertial Odometry with Visual Hallucination
[article]
2020
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Other Versions
Thermal cameras are commonly used for perception and inspection when the environment has low visibility. However, their use in odometry estimation is hampered by the lack of robust visual features. ...
However, in visually-denied scenarios (e.g. heavy smoke or darkness), pose estimates degrade or even fail. ...
As we have different (pseudo) ground truth format to compare with (VICON, Lidar Gmapping, or VINS-Mono), we align the predicted poses with the (pseudo) ground truth using Horn approaches and evaluate only ...
arXiv:1909.07231v2
fatcat:w4jtfllubvfsbl6a6lq3qnjxe4
Deep Learning-Based Monocular 3D Object Detection with Refinement of Depth Information
2022
Sensors
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
In contrast to LiDAR-based algorithms, the robustness of pseudo-LiDAR methods is still inferior. ...
Recently, the research on monocular 3D target detection based on pseudo-LiDAR data has made some progress. ...
Only the camera calibration matrix was used for projecting the depth map to pseudo-LiDAR. Our framework focuses on the BEV and 3D object detection. ...
doi:10.3390/s22072576
pmid:35408191
pmcid:PMC9003335
fatcat:t3oc4fcjfvgsxaxc2pemlebnle
ZoomNet: Part-Aware Adaptive Zooming Neural Network for 3D Object Detection
2020
PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Extensive experiments on the popular KITTI 3D detection dataset indicate ZoomNet surpasses all previous state-of-the-art methods by large margins (improved by 9.4% on APbv (IoU=0.7) over pseudo-LiDAR). ...
Though great progress has been made, challenges remain in estimating 3D pose for distant and occluded objects. ...
The concatenated left-right RoI features are used to regress 2D key-points and 3D dimensions. ...
doi:10.1609/aaai.v34i07.6945
fatcat:wngiiy3npvdxpnxvz5vzjassnm
ZoomNet: Part-Aware Adaptive Zooming Neural Network for 3D Object Detection
[article]
2020
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Extensive experiments on the popular KITTI 3D detection dataset indicate ZoomNet surpasses all previous state-of-the-art methods by large margins (improved by 9.4% on APbv (IoU=0.7) over pseudo-LiDAR). ...
Though great progress has been made, challenges remain in estimating 3D pose for distant and occluded objects. ...
The concatenated left-right RoI features are used to regress 2D key-points and 3D dimensions. ...
arXiv:2003.00529v1
fatcat:kfcz5eoqvnaflhhlfqt6mkeb7a
A Survey on Deep Learning Based Methods and Datasets for Monocular 3D Object Detection
2021
Electronics
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
Owing to recent advancements in deep learning methods and relevant databases, it is becoming increasingly easier to recognize 3D objects using only RGB images from single viewpoints. ...
For relatively low-cost data acquisition systems without depth sensors or cameras at multiple viewpoints, we first consider existing databases with 2D RGB photos and their relevant attributes. ...
Conversion of an estimated depth map Pseudo-LiDAR [23] Represent. ...
doi:10.3390/electronics10040517
fatcat:rziqhrkefvelpg3vgb6qxfprte
Object-Centric Stereo Matching for 3D Object Detection
[article]
2020
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Other Versions
Using stereo cameras to solve this task is a cost-effective alternative to the widely used LiDAR sensor. ...
Safe autonomous driving requires reliable 3D object detection-determining the 6 DoF pose and dimensions of objects of interest. ...
Existing methods can be categorized by the sensors they use: LiDAR [1] , [2] , [3] , [4] , [5] , LiDAR and camera [6] , [7] , [8] , monocular camera [9] , [10] , [11] , and stereo camera setups ...
arXiv:1909.07566v2
fatcat:pqlye2plqfbvvjoocnetvqy5d4
Hallucinating Dense Optical Flow from Sparse Lidar for Autonomous Vehicles
[article]
2018
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2019; you can also visit the original URL.
The file type is application/pdf.
To train this network we introduce a dataset with approximately 20K lidar samples of the Kitti dataset which we have augmented with a pseudo ground-truth image-based optical flow computed using FlowNet2 ...
We demonstrate the effectiveness of our approach on Kitti, and show that despite using the low-resolution and sparse measurements of the lidar, we can regress dense optical flow maps which are at par with ...
For that we use the MatConvNet [33] Deep Learning Framework. ...
arXiv:1808.10542v1
fatcat:3jwwwzmgobf4jlm5ceflqrcu7q
Monocular 3D Object Detection: An Extrinsic Parameter Free Approach
[article]
2021
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
Other Versions
To this end, we propose a novel method to capture camera pose to formulate the detector free from extrinsic perturbation. ...
Due to the lack of insight in industrial application, existing methods on open datasets neglect the camera pose information, which inevitably results in the detector being susceptible to camera extrinsic ...
Camera pose detection. For camera extrinsic parame-
ters regression study, we evaluate the angular errors of the 5. ...
arXiv:2106.15796v2
fatcat:gobjw2dp2ffifgzndlsfmkmiay
PointLoc: Deep Pose Regressor for LiDAR Point Cloud Localization
[article]
2021
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
Other Versions
However, LiDAR point clouds are unordered and unstructured making it difficult to apply traditional deep learning regression models for this task. ...
In this paper, we present a novel end-to-end learning-based LiDAR relocalization framework, termed PointLoc, which infers 6-DoF poses directly using only a single point cloud as input, without requiring ...
PROBLEM STATEMENT We design a DNN-based framework for performing deep global pose regression using point cloud data from a LiDAR sensor, which is LiDAR relocalization. ...
arXiv:2003.02392v3
fatcat:otfe45l3rza5nkashcwyf72ree
3D Pedestrian Detection in Farmland by Monocular RGB Image and Far-Infrared Sensing
2021
Remote Sensing
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
The Pseudo-LiDAR [27] employed the DORN [12] network to estimate the pixel depth, then converted the depth map to the pseudo point cloud using the corresponding geometric relationship. ...
[14] designed networks to respectively predict the depth of each pixel and the camera pose. ...
doi:10.3390/rs13152896
fatcat:bzhakednbrcvjhgnna7lnowteq
« Previous
Showing results 1 — 15 out of 582 results