Clustering-Based Representation Learning through Output Translation and Its Application to Remote--Sensing Images

TL;DR

This paper introduces a novel clustering-based self-supervised learning method for remote-sensing images that improves representation quality by evenly distributing samples while maintaining neighborhood relations, outperforming existing methods.

Contribution

The paper proposes a new clustering-based approach with a theoretical measure of discriminativeness and an algorithm for output translation to enhance representation learning in remote sensing images.

Findings

Our method achieves comparable or superior performance to state-of-the-art approaches.

It demonstrates robustness and computational efficiency across various RSI datasets.

Theoretical analysis explains the importance of even distribution for discriminative representations.

Abstract

In supervised deep learning, learning good representations for remote--sensing images (RSI) relies on manual annotations. However, in the area of remote sensing, it is hard to obtain huge amounts of labeled data. Recently, self--supervised learning shows its outstanding capability to learn representations of images, especially the methods of instance discrimination. Comparing methods of instance discrimination, clustering--based methods not only view the transformations of the same image as ``positive" samples but also similar images. In this paper, we propose a new clustering-based method for representation learning. We first introduce a quantity to measure representations' discriminativeness and from which we show that even distribution requires the most discriminative representations. This provides a theoretical insight into why evenly distributing the images works well. We notice that only the even distributions that preserve representations' neighborhood relations are desirable. Therefore, we develop an algorithm that translates the outputs of a neural network to achieve the goal of evenly distributing the samples while preserving outputs' neighborhood relations. Extensive experiments have demonstrated that our method can learn representations that are as good as or better than the state of the art approaches, and that our method performs computationally efficiently and robustly on various RSI datasets.