Artificial Generation of Big Data for Improving Image Classification: A Generative Adversarial Network Approach on SAR Data

Dimitrios Marmanis; Wei Yao; Fathalrahman Adam; Mihai Datcu; Peter Reinartz; Konrad Schindler; Jan Dirk Wegner; Uwe Stilla

Artificial Generation of Big Data for Improving Image Classification: A Generative Adversarial Network Approach on SAR Data

Computer Vision and Pattern Recognition 2017-11-07 v1

Authors: Dimitrios Marmanis , Wei Yao , Fathalrahman Adam , Mihai Datcu , Peter Reinartz , Konrad Schindler , Jan Dirk Wegner , Uwe Stilla

View on arXiv ↗ PDF ↗

Abstract

Very High Spatial Resolution (VHSR) large-scale SAR image databases are still an unresolved issue in the Remote Sensing field. In this work, we propose such a dataset and use it to explore patch-based classification in urban and periurban areas, considering 7 distinct semantic classes. In this context, we investigate the accuracy of large CNN classification models and pre-trained networks for SAR imaging systems. Furthermore, we propose a Generative Adversarial Network (GAN) for SAR image generation and test, whether the synthetic data can actually improve classification accuracy.

Keywords

generative adversarial network synthetic data generation image generation

Cite

@article{arxiv.1711.02010,
  title  = {Artificial Generation of Big Data for Improving Image Classification: A Generative Adversarial Network Approach on SAR Data},
  author = {Dimitrios Marmanis and Wei Yao and Fathalrahman Adam and Mihai Datcu and Peter Reinartz and Konrad Schindler and Jan Dirk Wegner and Uwe Stilla},
  journal= {arXiv preprint arXiv:1711.02010},
  year   = {2017}
}

Comments

Submitted for review in "Big Data from Space 2017" conference

Artificial Generation of Big Data for Improving Image Classification: A Generative Adversarial Network Approach on SAR Data

Abstract

Keywords

Cite

Comments

Related papers