Multi-Task Learning for Segmentation of Building Footprints with Deep Neural Networks

TL;DR

This paper introduces a cascaded multi-task loss for deep neural networks to improve boundary accuracy in high-resolution satellite image segmentation, outperforming existing methods.

Contribution

It proposes a novel multi-task loss function that enhances boundary preservation in semantic segmentation of satellite imagery.

Findings

Outperforms state-of-the-art methods by 8.3% on Inria dataset

Improves boundary accuracy without additional post-processing

Effective on large-scale high-resolution images

Abstract

The increased availability of high resolution satellite imagery allows to sense very detailed structures on the surface of our planet. Access to such information opens up new directions in the analysis of remote sensing imagery. However, at the same time this raises a set of new challenges for existing pixel-based prediction methods, such as semantic segmentation approaches. While deep neural networks have achieved significant advances in the semantic segmentation of high resolution images in the past, most of the existing approaches tend to produce predictions with poor boundaries. In this paper, we address the problem of preserving semantic segmentation boundaries in high resolution satellite imagery by introducing a new cascaded multi-task loss. We evaluate our approach on Inria Aerial Image Labeling Dataset which contains large-scale and high resolution images. Our results show that we are able to outperform state-of-the-art methods by 8.3\% without any additional post-processing step.