ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Download
Publications Copernicus
Download
Citation
Articles | Volume IV-2/W7
ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci., IV-2/W7, 167–173, 2019
https://doi.org/10.5194/isprs-annals-IV-2-W7-167-2019
ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci., IV-2/W7, 167–173, 2019
https://doi.org/10.5194/isprs-annals-IV-2-W7-167-2019

  16 Sep 2019

16 Sep 2019

A COMPARATIVE ANALYSIS OF UNSUPERVISED AND SEMI-SUPERVISED REPRESENTATION LEARNING FOR REMOTE SENSING IMAGE CATEGORIZATION

P. J. Soto1, J. D. Bermudez1, P. N. Happ1, and R. Q. Feitosa1,2 P. J. Soto et al.
  • 1Dept. of Electrical Engineering, Pontifical Catholic University of Rio de Janeiro, Brazil
  • 2Rio de Janeiro State University, Brazil

Keywords: Generative Adversarial Networks, Deep Learning, Semi-supervised Learning, Representation Learning

Abstract. This work aims at investigating unsupervised and semi-supervised representation learning methods based on generative adversarial networks for remote sensing scene classification. The work introduces a novel approach, which consists in a semi-supervised extension of a prior unsupervised method, known as MARTA-GAN. The proposed approach was compared experimentally with two baselines upon two public datasets, UC-MERCED and NWPU-RESISC45. The experiments assessed the performance of each approach under different amounts of labeled data. The impact of fine-tuning was also investigated. The proposed method delivered in our analysis the best overall accuracy under scarce labeled samples, both in terms of absolute value and in terms of variability across multiple runs.