How Useful is Region-based Classification of Remote Sensing Images in a Deep Learning Framework? - Archive ouverte HAL Access content directly
Conference Papers Year : 2016

How Useful is Region-based Classification of Remote Sensing Images in a Deep Learning Framework?

Abstract

In this paper, we investigate the impact of segmentation algorithms as a preprocessing step for classification of remote sensing images in a deep learning framework. Especially, we address the issue of segmenting the image into regions to be classified using pre-trained deep neural networks as feature extractors for an SVM-based classifier. An efficient segmentation as a preprocessing step helps learning by adding a spatially-coherent structure to the data. Therefore, we compare algorithms producing superpixels with more traditional remote sensing segmentation algorithms and measure the variation in terms of classification accuracy. We establish that superpixel algorithms allow for a better classification accuracy as a homogenous and compact segmentation favors better generalization of the training samples.
Fichier principal
Vignette du fichier
AudebertLeSauxLefevre_HowUseful.pdf (1.48 Mo) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-01320016 , version 1 (02-11-2016)

Identifiers

Cite

Nicolas Audebert, Bertrand Le Saux, Sébastien Lefèvre. How Useful is Region-based Classification of Remote Sensing Images in a Deep Learning Framework?. IEEE International Geosciences and Remote Sensing Symposium (IGARSS), Jul 2016, Beijing, China. ⟨10.1109/IGARSS.2016.7730327⟩. ⟨hal-01320016⟩
664 View
417 Download

Altmetric

Share

Gmail Facebook X LinkedIn More