Explainable autonomous driving systems (EADS) are emerging recently as a combinatory field of explainable artificial intelligence (XAI) and vehicular automation (VA). EADS explains events, ambient environments, and engine operations of an autonomous driving vehicular, and it also delivers explainable results in an orderly manner. Explainable semantic segmentation (ESS) plays an essential role in building EADS, where it offers visual attention that helps the drivers to be aware of the ambient objects irrespective if they are roads, pedestrians, animals, or other objects. In this paper, we propose the first ESS model for EADS based on the variation autoencoder (VAE), and it uses the multiscale first-order derivatives between the latent space and the encoder layers to capture the curvatures of the neurons’ responses. Our model is termed as Mgrad2VAE and is bench-marked on the SYNTHIA dataset, where it outperforms the recent deep models in terms of image segmentation metrics.

Towards explainable semantic segmentation for autonomous driving systems by multi-scale variational attention / M. Abukmeil, A. Genovese, V. Piuri, F. Rundo, F. Scotti - In: 2021 IEEE International Conference on Autonomous Systems (ICAS)[s.l] : IEEE, 2021. - ISBN 978-1-7281-7289-7. - pp. 1-5 (( Intervento presentato al 1. convegno ICAS 2021 tenutosi a Montreal nel 2021 [10.1109/ICAS49788.2021.9551172].

Towards explainable semantic segmentation for autonomous driving systems by multi-scale variational attention

M. Abukmeil;A. Genovese;V. Piuri;F. Scotti
2021

Abstract

Explainable autonomous driving systems (EADS) are emerging recently as a combinatory field of explainable artificial intelligence (XAI) and vehicular automation (VA). EADS explains events, ambient environments, and engine operations of an autonomous driving vehicular, and it also delivers explainable results in an orderly manner. Explainable semantic segmentation (ESS) plays an essential role in building EADS, where it offers visual attention that helps the drivers to be aware of the ambient objects irrespective if they are roads, pedestrians, animals, or other objects. In this paper, we propose the first ESS model for EADS based on the variation autoencoder (VAE), and it uses the multiscale first-order derivatives between the latent space and the encoder layers to capture the curvatures of the neurons’ responses. Our model is termed as Mgrad2VAE and is bench-marked on the SYNTHIA dataset, where it outperforms the recent deep models in terms of image segmentation metrics.
Autonomous Driving System; VAE; XAI; ESS
Settore INF/01 - Informatica
Settore ING-INF/05 - Sistemi di Elaborazione delle Informazioni
Multi-Owner data Sharing for Analytics and Integration respecting Confidentiality and Owner control (MOSAICrOWN)
Machine Learning-based, Networking and Computing Infrastructure Resource Management of 5G and beyond Intelligent Networks (MARSAL)
PRIN201719SDECA_01 - High quality Open data Publishing and Enrichment (HOPE) - DE CAPITANI DI VIMERCATI, SABRINA - PRIN2017 - PRIN bando 2017 - 2019
Book Part (author)
File in questo prodotto:
File Dimensione Formato  
2021_icas.pdf

accesso aperto

Tipologia: Post-print, accepted manuscript ecc. (versione accettata dall'editore)
Dimensione 6.04 MB
Formato Adobe PDF
6.04 MB Adobe PDF Visualizza/Apri
Towards_Explainable_Semantic_Segmentation_for_Autonomous_Driving_Systems_by_Multi-Scale_Variational_Attention.pdf

accesso riservato

Tipologia: Publisher's version/PDF
Dimensione 3.73 MB
Formato Adobe PDF
3.73 MB Adobe PDF   Visualizza/Apri   Richiedi una copia
Pubblicazioni consigliate

Caricamento pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/2434/845430
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 3
  • ???jsp.display-item.citation.isi??? ND
social impact