Unsupervised Domain Adaptation Using Generative Adversarial Networks for Semantic Segmentation of Aerial Images

Bilel Benjdira,Yakoub Bazi,Anis Koubaa,Kais Ouni

doi:10.3390/rs11111369

Abstract

Segmenting aerial images is of great potential in surveillance and scene understanding of urban areas. It provides a mean for automatic reporting of the different events that happen in inhabited areas. This remarkably promotes public safety and traffic management applications. After the wide adoption of convolutional neural networks methods, the accuracy of semantic segmentation algorithms could easily surpass 80% if a robust dataset is provided. Despite this success, the deployment of a pretrained segmentation model to survey a new city that is not included in the training set significantly decreases accuracy. This is due to the domain shift between the source dataset on which the model is trained and the new target domain of the new city images. In this paper, we address this issue and consider the challenge of domain adaptation in semantic segmentation of aerial images. We designed an algorithm that reduces the domain shift impact using generative adversarial networks (GANs). In the experiments, we tested the proposed methodology on the International Society for Photogrammetry and Remote Sensing (ISPRS) semantic segmentation dataset and found that our method improves overall accuracy from 35% to 52% when passing from the Potsdam domain (considered as source domain) to the Vaihingen domain (considered as target domain). In addition, the method allows efficiently recovering the inverted classes due to sensor variation. In particular, it improves the average segmentation accuracy of the inverted classes due to sensor variation from 14% to 61%.

Highlights

Semantic segmentation is an image analysis task that assigns for every pixel in an input image a label that describes the class of its enclosing region
The major contributions of our work can be presented as follows: (1) To the best of our knowledge, no previous works have addressed the problem of domain adaptation for semantic segmentation in aerial imagery using generative adversarial networks (GANs). (2) We demonstrated that our approach mitigates the domain shift problem for cross-domain semantic segmentation in aerial imagery, which allows the portability of the semantic segmentation model over different image domains
We have proposed a new method for domain adaptation in semantic segmentation of aerial imagery based on GANs

Summary

Introduction

Semantic segmentation is an image analysis task that assigns for every pixel in an input image a label that describes the class of its enclosing region. Semantic segmentation can be used in aerial imagery in a variety of potential applications, like urban area monitoring and planning, traffic management and analysis, hazard detection and avoidance, and so on. This potential is boosted by the increasing adoption of unmanned aerial vehicles (UAVs). UAVs make the surveillance of inhabited areas easier due to their flexibility, great mobility, and the high resolution images that they can gather and stream in real time These images can be automatically processed by accurate semantic segmentation algorithms to substantially reinforce the ability to analyze and describe the surveyed scenes automatically

Objectives

Methods

Results

Conclusion