Simple baseline for vehicle pose estimation: experimental validation
Authors
Corrales Sánchez, Héctor; Hernández Martínez, Antonio; Izquierdo Gonzalo, Rubén; Hernández Parra, Noelia; Parra Alonso, Ignacio; [et al.]Identifiers
Permanent link (URI): http://hdl.handle.net/10017/59532DOI: 10.1109/ACCESS.2020.3010307
ISSN: 2169-3536
Publisher
IEEE
Date
2020-07-20Affiliation
Universidad de Alcalá. Departamento de Automática; Universidad de Alcalá. Departamento de Ciencias de la ComputaciónFunders
Comunidad de Madrid
Agencia Estatal de Investigación
European Commission
Ministerio de Economía y Competitividad
Bibliographic citation
Corrales Sánchez, Héctor, Hernández Martínez, Antonio, Izquierdo Gonzalo, Rubén, Hernández Parra, Noelia, Fernández-Llorca, David, 2020, IEEE Access, v. 8, p. 132539-132550
Keywords
Vehicle pose estimation
Vehicle keypoints detection
CNNs
Heat maps
Human pose estimation
Experimental validation
Project
info:eu-repo/grantAgreement/CAM//S2018%2FEMT-4362/SEGURIDAD DE VEHÍCULOS PARA UNA MOVILIDAD INTELIGENTE, SOSTENIBLE, SEGURA E INTEGRADORA/SEGVAUTO 4.0-CM
info:eu-repo/grantAgreement/AEI/Plan Estatal de Investigación Científica y Técnica y de Innovación 2013-2016/DPI2017-90035-R/ES/INTERACCION PREDICTIVA ENTRE VEHICULOS AUTONOMOS COOPERATIVOS Y USUARIOS VULNERABLES DE CARRETERA ORIENTADA AL USUARIO FINAL/
info:eu-repo/grantAgreement/EC/H2020/723021/BRIDGING GAPS FOR THE ADOPTION OF AUTOMATED VEHICLES/BRAVE
info:eu-repo/grantAgreement/MINECO//PCIN-2017-086
info:eu-repo/grantAgreement/EC/H2020/737469//AutoDrive
Document type
info:eu-repo/semantics/article
Version
info:eu-repo/semantics/publishedVersion
Publisher's version
https://doi.org/10.1109/ACCESS.2020.3010307Rights
Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0)
Access rights
info:eu-repo/semantics/openAccess
Abstract
Significant progress on human and vehicle pose estimation has been achieved in recent years. The performance of these methods has evolved from poor to remarkable in just a couple of years. This improvement has been obtained from increasingly complex architectures. In this paper, we explore the applicability of simple baseline methods by adding a few deconvolutional layers on a backbone network to estimate heat maps that correspond to the vehicle keypoints. This approach has been proven to be very effective for human pose estimation. The results are analyzed on the PASCAL3DC dataset, achieving state-of-the-art results. In addition, a set of experiments has been conducted to study current shortcomings in vehicle keypoints labelling, which adversely affect performance. A new strategy for de ning vehicle keypoints is presented and validated with our customized dataset with extended keypoints.
Files in this item
Files | Size | Format |
|
---|---|---|---|
Simple_baseline_IEEE_Access_20 ... | 1.560Mb |
|
Files | Size | Format |
|
---|---|---|---|
Simple_baseline_IEEE_Access_20 ... | 1.560Mb |
|