Advanced graph deep learning for ...
Document type :
Communication dans un congrès avec actes
Permalink :
Title :
Advanced graph deep learning for High-dimensional image analysis: challenges and opportunities
Author(s) :
Hanachi, Refka [Auteur]
Université de la Manouba [Tunisie] [UMA]
Sellami, Akrem [Auteur]
Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189 [CRIStAL]
Farah, Imed Riadh [Auteur]
Université de la Manouba [Tunisie] [UMA]
Dalla Mura, Mauro [Auteur]
GIPSA - Signal Images Physique [GIPSA-SIGMAPHY]
Université de la Manouba [Tunisie] [UMA]
Sellami, Akrem [Auteur]
Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189 [CRIStAL]
Farah, Imed Riadh [Auteur]
Université de la Manouba [Tunisie] [UMA]
Dalla Mura, Mauro [Auteur]
GIPSA - Signal Images Physique [GIPSA-SIGMAPHY]
Conference title :
ATSIP 2024 - IEEE 7th International Conference on Advanced Technologies, Signal and Image Processing (ATSIP)
City :
Sousse
Country :
Tunisie
Start date of the conference :
2024-07-11
Publisher :
IEEE
English keyword(s) :
Graph deep learning
3D images
High-dimensional image analysis
Graph partitioning
Large-scale graph data
3D images
High-dimensional image analysis
Graph partitioning
Large-scale graph data
HAL domain(s) :
Sciences de l'ingénieur [physics]/Traitement du signal et de l'image [eess.SP]
English abstract : [en]
High-dimensional image analysis, such as Hyperspectral Imaging (HSI) data, poses unique challenges due to their high dimensionality and non-Euclidean structures, making their analysis and classification complex. In this ...
Show more >High-dimensional image analysis, such as Hyperspectral Imaging (HSI) data, poses unique challenges due to their high dimensionality and non-Euclidean structures, making their analysis and classification complex. In this study, we explore the use of both graph deep learning (GDL) and multi-view graph representation learning for HSI classification. Furthermore, we present our proposed approach of multi-view Graph Convolutional Networks (GCNs) and how it leverages multiple views of the data by combining spectral and spatial features to improve classification accuracy. We discuss then specific challenges encountered when training our model on large HSIs, including managing large-scale graph data. We also discuss promising opportunities to overcome these challenges. By highlighting the challenges and opportunities associated with GDL and multi-view GCN usage for HSI classification, this study aims to shed light on recent developments and future prospects in this rapidly evolving field.Show less >
Show more >High-dimensional image analysis, such as Hyperspectral Imaging (HSI) data, poses unique challenges due to their high dimensionality and non-Euclidean structures, making their analysis and classification complex. In this study, we explore the use of both graph deep learning (GDL) and multi-view graph representation learning for HSI classification. Furthermore, we present our proposed approach of multi-view Graph Convolutional Networks (GCNs) and how it leverages multiple views of the data by combining spectral and spatial features to improve classification accuracy. We discuss then specific challenges encountered when training our model on large HSIs, including managing large-scale graph data. We also discuss promising opportunities to overcome these challenges. By highlighting the challenges and opportunities associated with GDL and multi-view GCN usage for HSI classification, this study aims to shed light on recent developments and future prospects in this rapidly evolving field.Show less >
Language :
Anglais
Peer reviewed article :
Oui
Audience :
Internationale
Popular science :
Non
Collections :
Source :
Submission date :
2025-01-22T04:50:52Z