Exploiting GAN-Based SAR to Optical Image Transcoding for Improved Classification via Deep Learning

Conference: EUSAR 2018 - 12th European Conference on Synthetic Aperture Radar
06/04/2018 - 06/07/2018 at Aachen, Germany

Proceedings: EUSAR 2018

Pages: 6Language: englishTyp: PDF

Personal VDE Members are entitled to a 10% discount on this title

Ley, Andreas; Dhondt, Olivier; Valade, Sebastien; Haensch, Ronny; Hellwich, Olaf (Dep. Computer Vision & Remote Sensing, Technische Universit├Ąt Berlin, Germany)

Learning the proxy task of transcoding SAR images into optical images forces an employed conditional generative adversarial network (GAN) to distinguish between different land surfaces. Such a network can then be used to build a classifier with significantly fewer free parameters that generalizes well even when trained on a very small amount of labeled data. We train such a GAN on aligned Sentinel-1 and Sentinel-2 image pairs. We then show that a pre-trained classifier using these features learned from transcoding outperforms classifiers that are trained from scratch when only a very limited amount of labeled pixels are available for training.