Pose-Based Pedestrian Crossing Intention Prediction with Recurrent Graph Convolutions and Explainability-Driven Data Augmentation

This paper presents three contributions to the field of autonomous driving. Firstly, we present a Graph Neural Network designed for the pedestrian crossing intention prediction. The proposed architecture significantly outperforms the state-of-the-art methods in terms of f1-score, precision, and reca...

ver descrição completa

Detalhes bibliográficos
Autor: García Romera, Abel
Formato: tesis de maestría
Fecha de publicación:2023
País:España
Recursos:Universitat Autònoma de Barcelona
Repositorio:Dipòsit Digital de Documents de la UAB
Idioma:inglés
OAI Identifier:oai:ddd.uab.cat:293067
Acesso em linha:https://ddd.uab.cat/record/293067
Access Level:acceso abierto
Palavra-chave:Advanced Driver Assistance Systems
Data augmentation
Explainability
Graph Neural Networks
Pedestrian Crossing Intention
id ES_fd3d44041fb03d9783d5861e242fdace
oai_identifier_str oai:ddd.uab.cat:293067
network_acronym_str ES
network_name_str España
repository_id_str
spelling Pose-Based Pedestrian Crossing Intention Prediction with Recurrent Graph Convolutions and Explainability-Driven Data AugmentationGarcía Romera, AbelAdvanced Driver Assistance SystemsData augmentationExplainabilityGraph Neural NetworksPedestrian Crossing IntentionThis paper presents three contributions to the field of autonomous driving. Firstly, we present a Graph Neural Network designed for the pedestrian crossing intention prediction. The proposed architecture significantly outperforms the state-of-the-art methods in terms of f1-score, precision, and recall. The input consists of the skeletons of the pedestrians. It extracts features from the skeletons using a Recurrent Graph Convolutional Layer that captures both the spatial and temporal dependencies. Secondly, an explainability analysis has been performed using our implementation of Grad-Cam. Finally, based on the results of the explainability, we have proposed a data augmentation method that generates new skeletons.Aquest article presenta tres contribucions al camp de la conducció autònoma. Primer, presentem una Graph Neural Network dissenyada per a la predicció de la intenció de creuar dels vianants. L'arquitectura proposada supera significativament els mètodes actuals en termes de f1-score, precision i recall. L'entrada consisteix en els esquelets dels vianants. Extreu característiques dels esquelets mitjançant una Recurrent Graph Convolutional Layer que captura les dependències espacials i temporals. En segon lloc, s'ha realitzat una anàlisi d'explainability mitjançant la nostra implementació de Grad-Cam. Finalment, a partir dels resultats de l'explainability, hem proposat un mètode de data augmentation que genera nous esquelets.Universitat Autònoma de Barcelona. Escola d'EnginyeriaLópez Peña, Antonio M. 22023-01-0120232023-01-01Treball de fi de postgrauhttp://purl.org/coar/resource_type/c_bdccVoRhttp://purl.org/coar/version/c_970fb48d4fbd8a85info:eu-repo/semantics/masterThesisapplication/pdfhttps://ddd.uab.cat/record/293067reponame:Dipòsit Digital de Documents de la UABinstname:Universitat Autònoma de BarcelonaInglésengopen accesshttp://purl.org/coar/access_right/c_abf2Aquest document està subjecte a una llicència d'ús Creative Commons. Es permet la reproducció total o parcial, la distribució, la comunicació pública de l'obra, i la creació d'obres derivades, sempre que no sigui amb finalitats comercials i que es distribueixin sota la mateixa llicència que regula l'obra original. Cal que es reconegui l'autoria de l'obra original.https://creativecommons.org/licenses/by-nc-sa/4.0/info:eu-repo/semantics/openAccessoai:ddd.uab.cat:2930672026-06-06T12:50:31Z
dc.title.none.fl_str_mv Pose-Based Pedestrian Crossing Intention Prediction with Recurrent Graph Convolutions and Explainability-Driven Data Augmentation
title Pose-Based Pedestrian Crossing Intention Prediction with Recurrent Graph Convolutions and Explainability-Driven Data Augmentation
spellingShingle Pose-Based Pedestrian Crossing Intention Prediction with Recurrent Graph Convolutions and Explainability-Driven Data Augmentation
García Romera, Abel
Advanced Driver Assistance Systems
Data augmentation
Explainability
Graph Neural Networks
Pedestrian Crossing Intention
title_short Pose-Based Pedestrian Crossing Intention Prediction with Recurrent Graph Convolutions and Explainability-Driven Data Augmentation
title_full Pose-Based Pedestrian Crossing Intention Prediction with Recurrent Graph Convolutions and Explainability-Driven Data Augmentation
title_fullStr Pose-Based Pedestrian Crossing Intention Prediction with Recurrent Graph Convolutions and Explainability-Driven Data Augmentation
title_full_unstemmed Pose-Based Pedestrian Crossing Intention Prediction with Recurrent Graph Convolutions and Explainability-Driven Data Augmentation
title_sort Pose-Based Pedestrian Crossing Intention Prediction with Recurrent Graph Convolutions and Explainability-Driven Data Augmentation
dc.creator.none.fl_str_mv García Romera, Abel
author García Romera, Abel
author_facet García Romera, Abel
author_role author
dc.contributor.none.fl_str_mv Universitat Autònoma de Barcelona. Escola d'Enginyeria
López Peña, Antonio M.
dc.subject.none.fl_str_mv Advanced Driver Assistance Systems
Data augmentation
Explainability
Graph Neural Networks
Pedestrian Crossing Intention
topic Advanced Driver Assistance Systems
Data augmentation
Explainability
Graph Neural Networks
Pedestrian Crossing Intention
description This paper presents three contributions to the field of autonomous driving. Firstly, we present a Graph Neural Network designed for the pedestrian crossing intention prediction. The proposed architecture significantly outperforms the state-of-the-art methods in terms of f1-score, precision, and recall. The input consists of the skeletons of the pedestrians. It extracts features from the skeletons using a Recurrent Graph Convolutional Layer that captures both the spatial and temporal dependencies. Secondly, an explainability analysis has been performed using our implementation of Grad-Cam. Finally, based on the results of the explainability, we have proposed a data augmentation method that generates new skeletons.
publishDate 2023
dc.date.none.fl_str_mv 2
2023-01-01
2023
2023-01-01
dc.type.none.fl_str_mv Treball de fi de postgrau
http://purl.org/coar/resource_type/c_bdcc
VoR
http://purl.org/coar/version/c_970fb48d4fbd8a85
dc.type.openaire.fl_str_mv info:eu-repo/semantics/masterThesis
format masterThesis
dc.identifier.none.fl_str_mv https://ddd.uab.cat/record/293067
url https://ddd.uab.cat/record/293067
dc.language.none.fl_str_mv Inglés
eng
language_invalid_str_mv Inglés
language eng
dc.rights.none.fl_str_mv open access
http://purl.org/coar/access_right/c_abf2
https://creativecommons.org/licenses/by-nc-sa/4.0/
dc.rights.openaire.fl_str_mv info:eu-repo/semantics/openAccess
rights_invalid_str_mv open access
http://purl.org/coar/access_right/c_abf2
https://creativecommons.org/licenses/by-nc-sa/4.0/
eu_rights_str_mv openAccess
dc.format.none.fl_str_mv application/pdf
dc.source.none.fl_str_mv reponame:Dipòsit Digital de Documents de la UAB
instname:Universitat Autònoma de Barcelona
instname_str Universitat Autònoma de Barcelona
reponame_str Dipòsit Digital de Documents de la UAB
collection Dipòsit Digital de Documents de la UAB
repository.name.fl_str_mv
repository.mail.fl_str_mv
_version_ 1869425521733926912
score 15,223283

Registos relacionados