Centre de documentation
scientifique

Learning and adapting robust features for satellite image segmentation on heterogeneous data sets / Sina Ghassemi in IEEE Transactions on geoscience and remote sensing, vol 57 n° 9 (September 2019)  

Public [article]  inIEEE Transactions on geoscience and remote sensing > vol 57 n° 9 (September 2019) . - pp 6517 - 6529 Titre : Learning and adapting robust features for satellite image segmentation on heterogeneous data sets Type de document : Article/Communication Auteurs : Sina Ghassemi, Auteur ; Attilio Friandrotti, Auteur ; Gianluca Francini, Auteur ; Enrico Magli, Auteur Année de publication : 2019 Article en page(s) : pp 6517 - 6529 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique [Termes IGN] apprentissage profond [Termes IGN] chaîne de traitement [Termes IGN] classification par réseau neuronal convolutif [Termes IGN] coût [Termes IGN] données hétérogènes [Termes IGN] image binaire [Termes IGN] image satellite [Termes IGN] méthode robuste [Termes IGN] réseau neuronal convolutif [Termes IGN] segmentation binaire [Termes IGN] segmentation d'image [Termes IGN] test de performance Résumé : (auteur) This paper addresses the problem of training a deep neural network for satellite image segmentation so that it can be deployed over images whose statistics differ from those used for training. For example, in postdisaster damage assessment, the tight time constraints make it impractical to train a network from scratch for each image to be segmented. We propose a convolutional encoder–decoder network able to learn visual representations of increasing semantic level as its depth increases, allowing it to generalize over a wider range of satellite images. Then, we propose two additional methods to improve the network performance over each specific image to be segmented. First, we observe that updating the batch normalization layers’ statistics over the target image improves the network performance without human intervention. Second, we show that refining a trained network over a few samples of the image boosts the network performance with minimal human intervention. We evaluate our architecture over three data sets of satellite images, showing the state-of-the-art performance in binary segmentation of previously unseen images and competitive performance with respect to more complex techniques in a multiclass segmentation task. Numéro de notice : A2019-341 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1109/TGRS.2019.2906689 Date de publication en ligne : 17/04/2019 En ligne : https://doi.org/10.1109/TGRS.2019.2906689 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=93379 [article]