A Scalable Self-supervised Learner for Hyperspectral Image Classification

Weili Kong, Lin Qi, Baisen Liu, Jiaming Pei

Keywords: Hyperspectral Image Classification, Self-supervised, Auxiliary Task.

TL;DR: We designed a simple and scalable self-supervised leaner for hyperspectral image classification which supplement domain knowledge by adding appropriate auxiliary tasks.

Abstract:

Learning-based Hyperspectral image classification methods have achieved fantastic performance due to their superior ability to represent features at the cost that these methods are complex, inflexible and weak to generalize. Thus, we propose a simple and scalable pretrained model which can greatly accelerate convergence rate and improve the performance in downstream classification task.
Experiments shows the performance of our method achieved state-of-the-art performance with limited training samples.