Convolutional Neural Networks Meet Markov Random Fields for Semantic Segmentation of Remote Sensing Images - Archive ouverte HAL Accéder directement au contenu
Chapitre D'ouvrage Année : 2024

Convolutional Neural Networks Meet Markov Random Fields for Semantic Segmentation of Remote Sensing Images

Résumé

In this chapter, we have addressed the critical problem of the impact of spatially sparse ground truth on semantic segmentation maps obtained by fully convolutional networks by proposing a novel method for the semantic segmentation of remote sensing images based on CNNs, hierarchical PGMs, and decision tree ensembles. The rationale is to leverage the spatial modeling capabilities of hierarchical PGMs and the multiscale data representation extracted by FCNs to mitigate the impact of incomplete ground truth and obtain accurate classification results in scenarios where exhaustive ground truth does not exist. The integration was performed through a quadtree topology.
Fichier non déposé

Dates et versions

hal-04576855 , version 1 (15-05-2024)

Licence

Copyright (Tous droits réservés)

Identifiants

Citer

Martina Pastorino, Gabriele Moser, Sebastiano Serpico, Josiane Zerubia. Convolutional Neural Networks Meet Markov Random Fields for Semantic Segmentation of Remote Sensing Images. Signal and Image Processing for Remote Sensing, CRC Press, pp.205-227, 2024, 9781003382010. ⟨10.1201/9781003382010-14⟩. ⟨hal-04576855⟩
17 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Mastodon Facebook X LinkedIn More