Loading…

Feature Aware Re-weighting (FAR) in Bird’s Eye View for LiDAR-based 3D object detection in autonomous driving applications

3D object detection is a key element for the perception of autonomous vehicles. LiDAR sensors are commonly used to perceive the surrounding area, producing a sparse representation of the scene in the form of a point cloud. The current trend is to use deep learning neural network architectures that p...

Full description

Saved in:
Bibliographic Details
Published in:Robotics and autonomous systems 2024-05, Vol.175, p.104664, Article 104664
Main Authors: Zamanakos, Georgios, Tsochatzidis, Lazaros, Amanatiadis, Angelos, Pratikakis, Ioannis
Format: Article
Language:English
Subjects:
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
cited_by
cites cdi_FETCH-LOGICAL-c253t-4117ecb05bade8dc03be29b44fc475a5aa706b2285180ca5b95be20930a7a9a63
container_end_page
container_issue
container_start_page 104664
container_title Robotics and autonomous systems
container_volume 175
creator Zamanakos, Georgios
Tsochatzidis, Lazaros
Amanatiadis, Angelos
Pratikakis, Ioannis
description 3D object detection is a key element for the perception of autonomous vehicles. LiDAR sensors are commonly used to perceive the surrounding area, producing a sparse representation of the scene in the form of a point cloud. The current trend is to use deep learning neural network architectures that predict 3D bounding boxes. The vast majority of architectures process the LiDAR point cloud directly but, due to computation and memory constraints, at some point they compress the input to a 2D Bird’s Eye View (BEV) representation. In this work, we propose a novel 2D neural network architecture, namely the Feature Aware Re-weighting Network, for feature extraction in BEV using local context via an attention mechanism, to improve the 3D detection performance of LiDAR-based detectors. Extensive experiments on five state-of-the-art detectors and three benchmarking datasets, namely KITTI, Waymo and nuScenes, demonstrate the effectiveness of the proposed method in terms of both detection performance and minimal added computational burden. We release our code at https://github.com/grgzam/FAR. •Feature Aware Re-weighting (FAR) in BEV for LiDAR-based 3D object detection in autonomous driving applications.•Extracted local context in BEV via attention, is essential for an improved performance.•Modularity of FAR Network allows for adaptation in existing 3D object detectors.•FAR Network is evaluated on five SOTA detectors and three datasets, KITTI, Waymo and nuScenes.
doi_str_mv 10.1016/j.robot.2024.104664
format article
fullrecord <record><control><sourceid>elsevier_cross</sourceid><recordid>TN_cdi_crossref_primary_10_1016_j_robot_2024_104664</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><els_id>S0921889024000472</els_id><sourcerecordid>S0921889024000472</sourcerecordid><originalsourceid>FETCH-LOGICAL-c253t-4117ecb05bade8dc03be29b44fc475a5aa706b2285180ca5b95be20930a7a9a63</originalsourceid><addsrcrecordid>eNp9kMtKA0EQRRtRMEa_wE0vdTGxH_NcuIh5qBAQgoq7ph81sYdkOnRPEgIu_A1_zy9xxrh2UxeKey9VB6FLSgaU0PSmGninXDNghMXtJk7T-Aj1aJ6xKCv42zHqkYLRKM8LcorOQqgIITzJeA99TEE2Gw94uJPtnEO0A7t4b2y9wFfT4fwa2xrfWW--P78CnuwBv1rY4dJ5PLPj4TxSMoDBfIydqkA32EDTinV1F5SbxtVu5TYBG2-3Xalcr5dWy84RztFJKZcBLv60j16mk-fRQzR7un8cDWeRZglvopjSDLQiiZIGcqMJV8AKFceljrNEJlJmJFWM5QnNiZaJKpLWQApOZCYLmfI-4ode7V0IHkqx9nYl_V5QIjqAohK_AEUHUBwAtqnbQwra07YWvAjaQq3BWN--KIyz_-Z_AKlcfDM</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Feature Aware Re-weighting (FAR) in Bird’s Eye View for LiDAR-based 3D object detection in autonomous driving applications</title><source>Elsevier</source><creator>Zamanakos, Georgios ; Tsochatzidis, Lazaros ; Amanatiadis, Angelos ; Pratikakis, Ioannis</creator><creatorcontrib>Zamanakos, Georgios ; Tsochatzidis, Lazaros ; Amanatiadis, Angelos ; Pratikakis, Ioannis</creatorcontrib><description>3D object detection is a key element for the perception of autonomous vehicles. LiDAR sensors are commonly used to perceive the surrounding area, producing a sparse representation of the scene in the form of a point cloud. The current trend is to use deep learning neural network architectures that predict 3D bounding boxes. The vast majority of architectures process the LiDAR point cloud directly but, due to computation and memory constraints, at some point they compress the input to a 2D Bird’s Eye View (BEV) representation. In this work, we propose a novel 2D neural network architecture, namely the Feature Aware Re-weighting Network, for feature extraction in BEV using local context via an attention mechanism, to improve the 3D detection performance of LiDAR-based detectors. Extensive experiments on five state-of-the-art detectors and three benchmarking datasets, namely KITTI, Waymo and nuScenes, demonstrate the effectiveness of the proposed method in terms of both detection performance and minimal added computational burden. We release our code at https://github.com/grgzam/FAR. •Feature Aware Re-weighting (FAR) in BEV for LiDAR-based 3D object detection in autonomous driving applications.•Extracted local context in BEV via attention, is essential for an improved performance.•Modularity of FAR Network allows for adaptation in existing 3D object detectors.•FAR Network is evaluated on five SOTA detectors and three datasets, KITTI, Waymo and nuScenes.</description><identifier>ISSN: 0921-8890</identifier><identifier>EISSN: 1872-793X</identifier><identifier>DOI: 10.1016/j.robot.2024.104664</identifier><language>eng</language><publisher>Elsevier B.V</publisher><subject>3D object detection ; Autonomous driving ; Bird’s Eye View ; Deep learning ; LiDAR ; Point cloud</subject><ispartof>Robotics and autonomous systems, 2024-05, Vol.175, p.104664, Article 104664</ispartof><rights>2024 Elsevier B.V.</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><cites>FETCH-LOGICAL-c253t-4117ecb05bade8dc03be29b44fc475a5aa706b2285180ca5b95be20930a7a9a63</cites><orcidid>0000-0001-6084-414X</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>314,780,784,27924,27925</link.rule.ids></links><search><creatorcontrib>Zamanakos, Georgios</creatorcontrib><creatorcontrib>Tsochatzidis, Lazaros</creatorcontrib><creatorcontrib>Amanatiadis, Angelos</creatorcontrib><creatorcontrib>Pratikakis, Ioannis</creatorcontrib><title>Feature Aware Re-weighting (FAR) in Bird’s Eye View for LiDAR-based 3D object detection in autonomous driving applications</title><title>Robotics and autonomous systems</title><description>3D object detection is a key element for the perception of autonomous vehicles. LiDAR sensors are commonly used to perceive the surrounding area, producing a sparse representation of the scene in the form of a point cloud. The current trend is to use deep learning neural network architectures that predict 3D bounding boxes. The vast majority of architectures process the LiDAR point cloud directly but, due to computation and memory constraints, at some point they compress the input to a 2D Bird’s Eye View (BEV) representation. In this work, we propose a novel 2D neural network architecture, namely the Feature Aware Re-weighting Network, for feature extraction in BEV using local context via an attention mechanism, to improve the 3D detection performance of LiDAR-based detectors. Extensive experiments on five state-of-the-art detectors and three benchmarking datasets, namely KITTI, Waymo and nuScenes, demonstrate the effectiveness of the proposed method in terms of both detection performance and minimal added computational burden. We release our code at https://github.com/grgzam/FAR. •Feature Aware Re-weighting (FAR) in BEV for LiDAR-based 3D object detection in autonomous driving applications.•Extracted local context in BEV via attention, is essential for an improved performance.•Modularity of FAR Network allows for adaptation in existing 3D object detectors.•FAR Network is evaluated on five SOTA detectors and three datasets, KITTI, Waymo and nuScenes.</description><subject>3D object detection</subject><subject>Autonomous driving</subject><subject>Bird’s Eye View</subject><subject>Deep learning</subject><subject>LiDAR</subject><subject>Point cloud</subject><issn>0921-8890</issn><issn>1872-793X</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><recordid>eNp9kMtKA0EQRRtRMEa_wE0vdTGxH_NcuIh5qBAQgoq7ph81sYdkOnRPEgIu_A1_zy9xxrh2UxeKey9VB6FLSgaU0PSmGninXDNghMXtJk7T-Aj1aJ6xKCv42zHqkYLRKM8LcorOQqgIITzJeA99TEE2Gw94uJPtnEO0A7t4b2y9wFfT4fwa2xrfWW--P78CnuwBv1rY4dJ5PLPj4TxSMoDBfIydqkA32EDTinV1F5SbxtVu5TYBG2-3Xalcr5dWy84RztFJKZcBLv60j16mk-fRQzR7un8cDWeRZglvopjSDLQiiZIGcqMJV8AKFceljrNEJlJmJFWM5QnNiZaJKpLWQApOZCYLmfI-4ode7V0IHkqx9nYl_V5QIjqAohK_AEUHUBwAtqnbQwra07YWvAjaQq3BWN--KIyz_-Z_AKlcfDM</recordid><startdate>202405</startdate><enddate>202405</enddate><creator>Zamanakos, Georgios</creator><creator>Tsochatzidis, Lazaros</creator><creator>Amanatiadis, Angelos</creator><creator>Pratikakis, Ioannis</creator><general>Elsevier B.V</general><scope>AAYXX</scope><scope>CITATION</scope><orcidid>https://orcid.org/0000-0001-6084-414X</orcidid></search><sort><creationdate>202405</creationdate><title>Feature Aware Re-weighting (FAR) in Bird’s Eye View for LiDAR-based 3D object detection in autonomous driving applications</title><author>Zamanakos, Georgios ; Tsochatzidis, Lazaros ; Amanatiadis, Angelos ; Pratikakis, Ioannis</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c253t-4117ecb05bade8dc03be29b44fc475a5aa706b2285180ca5b95be20930a7a9a63</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>3D object detection</topic><topic>Autonomous driving</topic><topic>Bird’s Eye View</topic><topic>Deep learning</topic><topic>LiDAR</topic><topic>Point cloud</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Zamanakos, Georgios</creatorcontrib><creatorcontrib>Tsochatzidis, Lazaros</creatorcontrib><creatorcontrib>Amanatiadis, Angelos</creatorcontrib><creatorcontrib>Pratikakis, Ioannis</creatorcontrib><collection>CrossRef</collection><jtitle>Robotics and autonomous systems</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Zamanakos, Georgios</au><au>Tsochatzidis, Lazaros</au><au>Amanatiadis, Angelos</au><au>Pratikakis, Ioannis</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Feature Aware Re-weighting (FAR) in Bird’s Eye View for LiDAR-based 3D object detection in autonomous driving applications</atitle><jtitle>Robotics and autonomous systems</jtitle><date>2024-05</date><risdate>2024</risdate><volume>175</volume><spage>104664</spage><pages>104664-</pages><artnum>104664</artnum><issn>0921-8890</issn><eissn>1872-793X</eissn><abstract>3D object detection is a key element for the perception of autonomous vehicles. LiDAR sensors are commonly used to perceive the surrounding area, producing a sparse representation of the scene in the form of a point cloud. The current trend is to use deep learning neural network architectures that predict 3D bounding boxes. The vast majority of architectures process the LiDAR point cloud directly but, due to computation and memory constraints, at some point they compress the input to a 2D Bird’s Eye View (BEV) representation. In this work, we propose a novel 2D neural network architecture, namely the Feature Aware Re-weighting Network, for feature extraction in BEV using local context via an attention mechanism, to improve the 3D detection performance of LiDAR-based detectors. Extensive experiments on five state-of-the-art detectors and three benchmarking datasets, namely KITTI, Waymo and nuScenes, demonstrate the effectiveness of the proposed method in terms of both detection performance and minimal added computational burden. We release our code at https://github.com/grgzam/FAR. •Feature Aware Re-weighting (FAR) in BEV for LiDAR-based 3D object detection in autonomous driving applications.•Extracted local context in BEV via attention, is essential for an improved performance.•Modularity of FAR Network allows for adaptation in existing 3D object detectors.•FAR Network is evaluated on five SOTA detectors and three datasets, KITTI, Waymo and nuScenes.</abstract><pub>Elsevier B.V</pub><doi>10.1016/j.robot.2024.104664</doi><orcidid>https://orcid.org/0000-0001-6084-414X</orcidid></addata></record>
fulltext fulltext
identifier ISSN: 0921-8890
ispartof Robotics and autonomous systems, 2024-05, Vol.175, p.104664, Article 104664
issn 0921-8890
1872-793X
language eng
recordid cdi_crossref_primary_10_1016_j_robot_2024_104664
source Elsevier
subjects 3D object detection
Autonomous driving
Bird’s Eye View
Deep learning
LiDAR
Point cloud
title Feature Aware Re-weighting (FAR) in Bird’s Eye View for LiDAR-based 3D object detection in autonomous driving applications
url http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-07T15%3A41%3A40IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-elsevier_cross&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Feature%20Aware%20Re-weighting%20(FAR)%20in%20Bird%E2%80%99s%20Eye%20View%20for%20LiDAR-based%203D%20object%20detection%20in%20autonomous%20driving%20applications&rft.jtitle=Robotics%20and%20autonomous%20systems&rft.au=Zamanakos,%20Georgios&rft.date=2024-05&rft.volume=175&rft.spage=104664&rft.pages=104664-&rft.artnum=104664&rft.issn=0921-8890&rft.eissn=1872-793X&rft_id=info:doi/10.1016/j.robot.2024.104664&rft_dat=%3Celsevier_cross%3ES0921889024000472%3C/elsevier_cross%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-c253t-4117ecb05bade8dc03be29b44fc475a5aa706b2285180ca5b95be20930a7a9a63%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true