Domain adaptation in remote sensing through cross-image synthesis with dictionaries

Giona Matasci, Frank De Morsier, Mikhail Kanevski, Devis Tuia

Research output: Chapter in Book/Report/Conference proceedingConference paper

Abstract

This contribution studies an approach based on dictionary learning which enables the alignment of the sparse representations of two images. Set in a domain adaptation context, the purpose of this work is to re-synthesize the pixels of a remote sensing image so that, for a given land-cover class, the new values of the samples are comparable across acquisitions. Consequently, the data space of a given source image can be converted to that of a related target image, or vice-versa. After the mentioned transformation, the performance of a classifier trained on the source image and used to predict the thematic classes on the target image is expected to be more robust. A linear transformation is derived thanks to an algorithm simultaneously learning the image-specific dictionaries and the mapping function bridging them via their respective sparse codes. Experiments on knowledge transfer among two co-registered VHR images acquired with different off-nadir angles show promising results. An appropriate cross-image synthesis yields an increased land-cover model portability from one acquisition to another.

Original languageEnglish
Title of host publicationInternational Geoscience and Remote Sensing Symposium (IGARSS)
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages3714-3717
Number of pages4
ISBN (Print)9781479957750
DOIs
Publication statusPublished - 4 Nov 2014
Externally publishedYes
EventJoint 2014 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2014 and the 35th Canadian Symposium on Remote Sensing, CSRS 2014 - Quebec City, Canada
Duration: 13 Jul 201418 Jul 2014

Publication series

NameInternational Geoscience and Remote Sensing Symposium (IGARSS)

Conference

ConferenceJoint 2014 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2014 and the 35th Canadian Symposium on Remote Sensing, CSRS 2014
CountryCanada
CityQuebec City
Period13/07/1418/07/14

Keywords

  • dataset shift
  • dictionary learning
  • image classification
  • sparse representation

Fingerprint Dive into the research topics of 'Domain adaptation in remote sensing through cross-image synthesis with dictionaries'. Together they form a unique fingerprint.

  • Cite this

    Matasci, G., De Morsier, F., Kanevski, M., & Tuia, D. (2014). Domain adaptation in remote sensing through cross-image synthesis with dictionaries. In International Geoscience and Remote Sensing Symposium (IGARSS) (pp. 3714-3717). [6947290] (International Geoscience and Remote Sensing Symposium (IGARSS)). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/IGARSS.2014.6947290