Rare Wildlife Recognition with Self-Supervised Representation Learning

TL;DR

This paper introduces a self-supervised learning approach using contrastive methods and data augmentation to improve wildlife recognition in aerial images, significantly reducing the need for annotated data while maintaining high accuracy.

Contribution

It demonstrates that combining contrastive self-supervised learning with data augmentation outperforms traditional supervised models, enabling effective wildlife recognition with minimal labeled data.

Findings

Outperforms ImageNet-pretrained models by a large margin

Maintains high recall with only 10% of training data

Combines contrastive learning with mixup for robust representations

Abstract

Automated animal censuses with aerial imagery are a vital ingredient towards wildlife conservation. Recent models are generally based on supervised learning and thus require vast amounts of training data. Due to their scarcity and minuscule size, annotating animals in aerial imagery is a highly tedious process. In this project, we present a methodology to reduce the amount of required training data by resorting to self-supervised pretraining. In detail, we examine a combination of recent contrastive learning methodologies like Momentum Contrast (MoCo) and Cross-Level Instance-Group Discrimination (CLD) to condition our model on the aerial images without the requirement for labels. We show that a combination of MoCo, CLD, and geometric augmentations outperforms conventional models pretrained on ImageNet by a large margin. Meanwhile, strategies for smoothing label or prediction distribution in supervised learning have been proven useful in preventing the model from overfitting. We combine the self-supervised contrastive models with image mixup strategies and find that it is useful for learning more robust visual representations. Crucially, our methods still yield favorable results even if we reduce the number of training animals to just 10%, at which point our best model scores double the recall of the baseline at similar precision. This effectively allows reducing the number of required annotations to a fraction while still being able to train high-accuracy models in such highly challenging settings.