Explaining Bounding Boxes in Deep Object Detectors Using Post Hoc Methods for Autonomous Driving Systems
Sensors · 14.01.2024 · 24 (2) · 516
DOI: 10.3390/s24020516 ↗Kurz gefasst
Fachartikel in der Zeitschrift Sensors, online erschienen am 14.01.2024. Verfasst von Caio Nogueira und 3 weiteren Autor:innen. Beteiligte Einrichtung: Faculty of Engineering, University of Porto, 4200-465 Porto, Portugal. Die Arbeit wurde bisher 12-mal zitiert (Crossref) und nennt 28 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Portuguese National Funds, through AICEP, E. P. E. and the Innovation and Digital Transition Program (COMPETE2030), Portuguese National Funds, Innovation and Digital Transition Program.
Abstract (Originalsprache)
Deep learning has rapidly increased in popularity, leading to the development of perception solutions for autonomous driving. The latter field leverages techniques developed for computer vision in other domains for accomplishing perception tasks such as object detection. However, the black-box nature of deep neural models and the complexity of the autonomous driving context motivates the study of explainability in these models that perform perception tasks. Moreover, this work explores explainable AI techniques for the object detection task in the context of autonomous driving. An extensive and detailed comparison is carried out between gradient-based and perturbation-based methods (e.g., D-RISE). Moreover, several experimental setups are used with different backbone architectures and different datasets to observe the influence of these aspects in the explanations. All the techniques explored consist of saliency methods, making their interpretation and evaluation primarily visual. Nevertheless, numerical assessment methods are also used. Overall, D-RISE and guided backpropagation obtain more localized explanations. However, D-RISE highlights more meaningful regions, providing more human-understandable explanations. To the best of our knowledge, this is the first approach to obtaining explanations focusing on the regression of the bounding box coordinates.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Faculty of Engineering, University of Porto, 4200-465 Porto, Portugal.
- Instituto de Engenharia de Sistemas e Computadores, Tecnologia e Ciência, 4200-465 Porto, Portugal.
ORCID-Profile: Caio Nogueira, Luís Fernandes, João N D Fernandes, Jaime S Cardoso
Förderung
- Portuguese National Funds, through AICEP, E. P. E. and the Innovation and Digital Transition Program (COMPETE2030)
- Portuguese National Funds
- Innovation and Digital Transition Program
- ATLAS—Trusted Autonomous Navigation
Zitieren
APA
Nogueira, C., Fernandes, L., Fernandes, J. N. D., & Cardoso, J. S. (2024). Explaining Bounding Boxes in Deep Object Detectors Using Post Hoc Methods for Autonomous Driving Systems. Sensors, 24(2), 516. https://doi.org/10.3390/s24020516
Vancouver
Nogueira C, Fernandes L, Fernandes JND, Cardoso JS. Explaining Bounding Boxes in Deep Object Detectors Using Post Hoc Methods for Autonomous Driving Systems. Sensors. 2024;24(2):516. doi:10.3390/s24020516
BibTeX
@article{Nogueira2024,
author = {Nogueira, Caio and Fernandes, Luís and Fernandes, João N D and Cardoso, Jaime S},
title = {Explaining Bounding Boxes in Deep Object Detectors Using Post Hoc Methods for Autonomous Driving Systems},
journal = {Sensors},
year = {2024},
volume = {24},
number = {2},
pages = {516},
doi = {10.3390/s24020516}
}
EUVIMED