Regularization of Building Boundaries in Satellite Images Using Adversarial and Regularized Losses

Stefano Zorzi, Friedrich Fraundorfer

Research output: Chapter in Book/Report/Conference proceedingConference paperpeer-review

Abstract

In this paper we present a method for building boundary re-finement and regularization in satellite images using a fullyconvolutional neural network trained with a combination ofadversarial and regularized losses. Compared to a pure MaskR-CNN model, the overall algorithm can achieve equivalentperformance in terms of accuracy and completeness. How-ever, unlike Mask R-CNN that produces irregular footprints,our framework generates regularized and visually pleasingbuilding boundaries which are beneficial in many applica-tions.
Original languageEnglish
Title of host publicationIGARSS 2019
PublisherIEEE Publications
Pages5140-5143
Number of pages4
Publication statusPublished - 1 Jul 2019

Keywords

  • Generative adversarial networks
  • build-ing segmentation
  • boundary refinement
  • satellite images

Cite this