Improving Contrastive Learning on Visually Homogeneous Mars Rover Images

TL;DR

This paper enhances contrastive learning for Mars rover images by addressing the issue of false negatives due to low visual diversity, using clustering and domain mixing to improve unsupervised feature learning and classification accuracy.

Contribution

It introduces two unsupervised methods—clustering and domain mixing—to reduce false negatives in contrastive learning on homogeneous Mars images, improving downstream classification.

Findings

Achieved a 3.06% increase in classification accuracy with only 10% labeled data.

Demonstrated that reducing false negatives improves contrastive learning performance.

Proved that unsupervised domain mixing enhances feature diversity and model robustness.

Abstract

Contrastive learning has recently demonstrated superior performance to supervised learning, despite requiring no training labels. We explore how contrastive learning can be applied to hundreds of thousands of unlabeled Mars terrain images, collected from the Mars rovers Curiosity and Perseverance, and from the Mars Reconnaissance Orbiter. Such methods are appealing since the vast majority of Mars images are unlabeled as manual annotation is labor intensive and requires extensive domain knowledge. Contrastive learning, however, assumes that any given pair of distinct images contain distinct semantic content. This is an issue for Mars image datasets, as any two pairs of Mars images are far more likely to be semantically similar due to the lack of visual diversity on the planet's surface. Making the assumption that pairs of images will be in visual contrast - when they are in fact not - results in pairs that are falsely considered as negatives, impacting training performance. In this study, we propose two approaches to resolve this: 1) an unsupervised deep clustering step on the Mars datasets, which identifies clusters of images containing similar semantic content and corrects false negative errors during training, and 2) a simple approach which mixes data from different domains to increase visual diversity of the total training dataset. Both cases reduce the rate of false negative pairs, thus minimizing the rate in which the model is incorrectly penalized during contrastive training. These modified approaches remain fully unsupervised end-to-end. To evaluate their performance, we add a single linear layer trained to generate class predictions based on these contrastively-learned features and demonstrate increased performance compared to supervised models; observing an improvement in classification accuracy of 3.06% using only 10% of the labeled data.