@Article{SotoVegaCoFeOrAlHeRo:2021:UnDoAd,
author = "Soto Vega, Pedro Juan and Costa, Gilson Alexandre Ostwald Pedro da
and Feitosa, Raul Queiroz and Ortega Adarme, Mabel Ximena and
Almeida, Cl{\'a}udio Aparecido de and Heipke, Christian and
Rottensteiner, Franz",
affiliation = "{Universidade Federal do Rio de Janeiro (UFRJ)} and {Universidade
do Estado do Rio de Janeiro (UERJ)} and {Universidade Federal do
Rio de Janeiro (UFRJ)} and {Universidade Federal do Rio de Janeiro
(UFRJ)} and {Instituto Nacional de Pesquisas Espaciais (INPE)} and
{Leibniz Universitat Hannover (LUH)} and {Leibniz Universitat
Hannover (LUH)}",
title = "An unsupervised domain adaptation approach for change detection
and its application to deforestation mapping in tropical biomes",
journal = "ISPRS Journal of Photogrammetry and Remote Sensing",
year = "2021",
volume = "181",
pages = "113--128",
month = "Nov.",
keywords = "Change detection, CycleGAN, Deep learning, Deforestation
detection, Domain adaptation, Remote sensing.",
abstract = "Changes in environmental conditions, geographical variability and
different sensor properties typically make it almost impossible to
employ previously trained classifiers for new data without a
significant drop in classification accuracy. Domain adaptation
(DA) techniques been proven useful to alleviate that problem. In
particular, appearance adaptation techniques may be used to adapt
images from a specific dataset in such a way that the generated
images have a style that is similar to the images from another
dataset. Such techniques are, however, prone to creating artifacts
that hinder proper classification of the adapted images. In this
work we propose an unsupervised DA approach for change detection
tasks, which is based on a particular appearance adaptation
method: the Cycle-Consistent Generative Adversarial Network
(CycleGAN). Specifically, we extend that method by introducing
additional constraints in the training phase of the model
components, which make it preserve the semantic structure and
class transitions in the adapted images. We evaluate the proposed
approach on a deforestation detection application, considering
different sites in the Amazon rain-forest and in the Brazilian
Cerrado (savanna) using Landsat-8 images. In the experiments, each
site corresponds to a domain, and the accuracy of a classifier
trained with images and references from one (source) domain is
measured in the classification of another (target) domain. The
results show that the proposed approach is successful in producing
artifact-free adapted images, which can be satisfactory classified
by the pre-trained source classifiers. On average, the accuracies
achieved in the classification of the adapted images outperformed
the baselines (when no adaptation was made) by 7.1% in terms of
mean average precision, and 9.1% in terms of F1-Score. To the best
of our knowledge, the proposed method is the first unsupervised
domain adaptation approach devised for change detection.",
doi = "10.1016/j.isprsjprs.2021.08.026",
url = "http://dx.doi.org/10.1016/j.isprsjprs.2021.08.026",
issn = "0924-2716",
language = "en",
targetfile = "vega_unsupervised.pdf",
urlaccessdate = "12 maio 2024"
}