Satellite Image Segmentation with Deep Residual Architectures for Time-Critical Applications

We address the problem of training a convolutional neural network for satellite images segmentation in emergency situations, where response time constraints prevent training the network from scratch. Such case is particularly challenging due to the large intra-class statistics variations between tra...

Full description

Saved in:
Bibliographic Details
Published in2018 26th European Signal Processing Conference (EUSIPCO) pp. 2235 - 2239
Main Authors Ghassemi, Sina, Sandu, Constantin, Fiandrotti, Attilio, Tonolo, Fabio Giulio, Boccardo, Piero, Francini, Gianluca, Magli, Enrico
Format Conference Proceeding
LanguageEnglish
Published EURASIP 01.09.2018
Subjects
Convolution
Decoding
Feature extraction
Image segmentation
Network architecture
Satellites
Training
Online AccessGet full text

Cover

More Information
Summary:We address the problem of training a convolutional neural network for satellite images segmentation in emergency situations, where response time constraints prevent training the network from scratch. Such case is particularly challenging due to the large intra-class statistics variations between training images and images to be segmented captured at different locations by different sensors. We propose a convolutional encoder-decoder network architecture where the encoder builds upon a residual architecture. We show that our proposed architecture enables learning features suitable to generalize the learning process across images with different statistics. Our architecture can accurately segment images that have no reference in the training set, whereas a minimal refinement of the trained network significantly boosts the segmentation accuracy.
ISSN:2076-1465
DOI:10.23919/EUSIPCO.2018.8553545