Self-supervised Contrastive Learning for Cross-domain Hyperspectral Image Representation

TL;DR

This paper presents a self-supervised contrastive learning framework for hyperspectral images that learns cross-domain representations without labels, improving classification performance over traditional methods.

Contribution

It introduces a novel cross-domain CNN architecture with contrastive learning for hyperspectral images, enabling effective representation learning without pixel-level annotations.

Findings

Self-supervised representations outperform models trained from scratch.

Contrastive learning effectively clusters spectral vectors within and across images.

The framework improves downstream classification accuracy.

Abstract

Recently, self-supervised learning has attracted attention due to its remarkable ability to acquire meaningful representations for classification tasks without using semantic labels. This paper introduces a self-supervised learning framework suitable for hyperspectral images that are inherently challenging to annotate. The proposed framework architecture leverages cross-domain CNN, allowing for learning representations from different hyperspectral images with varying spectral characteristics and no pixel-level annotation. In the framework, cross-domain representations are learned via contrastive learning where neighboring spectral vectors in the same image are clustered together in a common representation space encompassing multiple hyperspectral images. In contrast, spectral vectors in different hyperspectral images are separated into distinct clusters in the space. To verify that the learned representation through contrastive learning is effectively transferred into a downstream task, we perform a classification task on hyperspectral images. The experimental results demonstrate the advantage of the proposed self-supervised representation over models trained from scratch or other transfer learning methods.