Cycle-Consistent Adversarial Networks for Realistic Pervasive Change\n Generation in Remote Sensing Imagery

2019/11/28 by Christopher X. Ren, Amanda Ziemann, Ren, Christopher X. +5
Chemistry · Engineering · Environmental Science · #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #FOS: Electrical engineering #Image and Video Processing (eess.IV) #Machine Learning (cs.LG) #Remote Sensing in Agriculture #Remote-Sensing Image Classification #Spectroscopy and Chemometric Analyses #electronic engineering #information engineering

paper · pdf · doi:10.48550/arxiv.1911.12546

openalex publication_date 2019/11/28 · openalex created_date 2022/07/26 · openalex updated_date 2026/07/28

Abstract

This paper introduces a new method of generating realistic pervasive changes\nin the context of evaluating the effectiveness of change detection algorithms\nin controlled settings. The method, a cycle-consistent adversarial network\n(CycleGAN), requires low quantities of training data to generate realistic\nchanges. Here we show an application of CycleGAN in creating realistic\nsnow-covered scenes of multispectral Sentinel-2 imagery, and demonstrate how\nthese images can be used as a test bed for anomalous change detection\nalgorithms.\n

Related

• S2-cGAN: Self-Supervised Adversarial Representation Learning for Binary\n Change Detection in Multispectral Images
• CD-GAN: a robust fusion-based generative adversarial network for unsupervised remote sensing change detection with heterogeneous sensors
• Consistency Change Detection Framework for Unsupervised Remote Sensing Change Detection
• Changen2: Multi-Temporal Remote Sensing Generative Change Foundation Model
• A Dual Attentive Generative Adversarial Network for Remote Sensing Image Change Detection