Geography-Aware Self-Supervised Learning

TL;DR

This paper introduces a novel geography-aware self-supervised learning approach that leverages spatio-temporal data structures in remote sensing to improve image classification, detection, and segmentation performance, narrowing the gap with supervised methods.

Contribution

It proposes new training techniques exploiting spatial and temporal information in geo-located datasets to enhance contrastive learning for remote sensing applications.

Findings

Improves remote sensing image classification, detection, and segmentation accuracy.

Closes the performance gap between contrastive and supervised learning.

Enhances downstream tasks on geo-tagged ImageNet images.

Abstract

Contrastive learning methods have significantly narrowed the gap between supervised and unsupervised learning on computer vision tasks. In this paper, we explore their application to geo-located datasets, e.g. remote sensing, where unlabeled data is often abundant but labeled data is scarce. We first show that due to their different characteristics, a non-trivial gap persists between contrastive and supervised learning on standard benchmarks. To close the gap, we propose novel training methods that exploit the spatio-temporal structure of remote sensing data. We leverage spatially aligned images over time to construct temporal positive pairs in contrastive learning and geo-location to design pre-text tasks. Our experiments show that our proposed method closes the gap between contrastive and supervised learning on image classification, object detection and semantic segmentation for remote sensing. Moreover, we demonstrate that the proposed method can also be applied to geo-tagged ImageNet images, improving downstream performance on various tasks. Project Webpage can be found at this link geography-aware-ssl.github.io.