Active Label Refinement for Semantic Segmentation of Satellite Images

TL;DR

This paper introduces an active label refinement approach that improves semantic segmentation of satellite images by efficiently correcting initial labels obtained through low-cost methods, enhancing model accuracy.

Contribution

It proposes a novel active learning-based label refinement method tailored for satellite image segmentation, addressing label noise from low-cost labeling sources.

Findings

Active label refinement improves segmentation accuracy.

Active learning strategies reduce labeling costs.

Refined labels lead to better model performance.

Abstract

Remote sensing through semantic segmentation of satellite images contributes to the understanding and utilisation of the earth's surface. For this purpose, semantic segmentation networks are typically trained on large sets of labelled satellite images. However, obtaining expert labels for these images is costly. Therefore, we propose to rely on a low-cost approach, e.g. crowdsourcing or pretrained networks, to label the images in the first step. Since these initial labels are partially erroneous, we use active learning strategies to cost-efficiently refine the labels in the second step. We evaluate the active learning strategies using satellite images of Bengaluru in India, labelled with land cover and land use labels. Our experimental results suggest that an active label refinement to improve the semantic segmentation network's performance is beneficial.