Unpublished conference/Abstract (Scientific congresses and symposiums)
Physically Interpretable Probabilistic Domain Characterization
Halin, Anaïs; Pierard, Sébastien; Vandeghen, Renaud et al.
2024First International Workshop on AI-based All-Weather Surveillance System (AWSS)
Peer reviewed
 

Files


Full Text
Halin2024Physically-PresentationAWSS.pdf
Author postprint (30.18 MB)
Download

All documents in ORBi are protected by a user license.

Send to



Details



Keywords :
Domain Characterization; Domain Adaptation; Normalizing Flows; Autonomous Vehicles; Weather Prediction
Abstract :
[en] Characterizing domains is essential for models analyzing dynamic environments, as it allows them to adapt to evolving conditions or to hand the task over to backup systems when facing conditions outside their operational domain. Existing solutions typically characterize a domain by solving a regression or classification problem, which limits their applicability as they only provide a limited summarized description of the domain. In this paper, we present a novel approach to domain characterization by characterizing domains as probability distributions. Particularly, we develop a method to predict the likelihood of different weather conditions from images captured by vehicle-mounted cameras by estimating distributions of physical parameters using normalizing flows. To validate our proposed approach, we conduct experiments within the context of autonomous vehicles, focusing on predicting the distribution of weather parameters to characterize the operational domain. This domain is characterized by physical parameters (absolute characterization) and arbitrarily predefined domains (relative characterization). Finally, we evaluate whether a system can safely operate in a target domain by comparing it to multiple source domains where safety has already been established. This approach holds significant potential, as accurate weather prediction and effective domain adaptation are crucial for autonomous systems to adjust to dynamic environmental conditions.
Disciplines :
Electrical & electronics engineering
Author, co-author :
Halin, Anaïs   ;  Université de Liège - ULiège > Montefiore Institute of Electrical Engineering and Computer Science
Pierard, Sébastien   ;  Université de Liège - ULiège > Montefiore Institute of Electrical Engineering and Computer Science
Vandeghen, Renaud ;  Université de Liège - ULiège > Montefiore Institute of Electrical Engineering and Computer Science
Gérin, Benoît;  UCL - Université Catholique de Louvain
Zanella, Maxime;  UCL - Université Catholique de Louvain
Colot, Martin;  ULB - Université Libre de Bruxelles
Held, Jan ;  Université de Liège - ULiège > Montefiore Institute of Electrical Engineering and Computer Science
Cioppa, Anthony  ;  Université de Liège - ULiège > Montefiore Institute of Electrical Engineering and Computer Science
Jean, Emmanuel;  Multitel research and technological innovation centre
Bontempi, Gianluca;  ULB - Université Libre de Bruxelles
Mahmoudi, Saïd;  UMONS - Université de Mons
Macq, Benoît;  UCL - Université Catholique de Louvain
Van Droogenbroeck, Marc  ;  Université de Liège - ULiège > Département d'électricité, électronique et informatique (Institut Montefiore) > Télécommunications
More authors (3 more) Less
 These authors have contributed equally to this work.
Language :
English
Title :
Physically Interpretable Probabilistic Domain Characterization
Publication date :
09 December 2024
Event name :
First International Workshop on AI-based All-Weather Surveillance System (AWSS)
Event organizer :
Asian Conference on Computer Vision Workshops (ACCVW)
Event place :
Hanoï, Vietnam
Event date :
9 décembre 2024
Audience :
International
Peer reviewed :
Peer reviewed
Available on ORBi :
since 19 January 2025

Statistics


Number of views
6 (4 by ULiège)
Number of downloads
4 (2 by ULiège)

Bibliography


Similar publications



Contact ORBi