Publication

Back to overview

Land cover mapping at very high resolution with rotation equivariant CNNs: Towards small yet accurate models

Type of publication Peer-reviewed
Publikationsform Original article (peer-reviewed)
Author Marcos Diego, Marcos Diego, Volpi Michele, Kellenberger Benjamin, Tuia Devis, Tuia Devis,
Project Multimodal machine learning for remote sensing information fusion
Show all

Original article (peer-reviewed)

Journal ISPRS Journal of Photogrammetry and Remote Sensing
Volume (Issue) 145
Page(s) 96 - 107
Title of proceedings ISPRS Journal of Photogrammetry and Remote Sensing
DOI 10.1016/j.isprsjprs.2018.01.021

Open Access

URL https://arxiv.org/abs/1803.06253
Type of Open Access Repository (Green Open Access)

Abstract

© 2018 International Society for Photogrammetry and Remote Sensing, Inc. (ISPRS) In remote sensing images, the absolute orientation of objects is arbitrary. Depending on an object's orientation and on a sensor's flight path, objects of the same semantic class can be observed in different orientations in the same image. Equivariance to rotation, in this context understood as responding with a rotated semantic label map when subject to a rotation of the input image, is therefore a very desirable feature, in particular for high capacity models, such as Convolutional Neural Networks (CNNs). If rotation equivariance is encoded in the network, the model is confronted with a simpler task and does not need to learn specific (and redundant) weights to address rotated versions of the same object class. In this work we propose a CNN architecture called Rotation Equivariant Vector Field Network (RotEqNet) to encode rotation equivariance in the network itself. By using rotating convolutions as building blocks and passing only the values corresponding to the maximally activating orientation throughout the network in the form of orientation encoding vector fields, RotEqNet treats rotated versions of the same object with the same filter bank and therefore achieves state-of-the-art performances even when using very small architectures trained from scratch. We test RotEqNet in two challenging sub-decimeter resolution semantic labeling problems, and show that we can perform better than a standard CNN while requiring one order of magnitude less parameters.
-