Contrastive Learning for Regression on Hyperspectral Data

The use of such approach on hyperspectral data is still encountering some challenges especially regarding the Contrastive learning has demonstrated great effectiveness in augmentation techniques to be used. Data augmentation techniques representation learning especially for image classification often used for general images (e.g., image rotation) tasks. However, there is still a shortage in the studies targeting are not applicable to hyperspectral data. In this article, we investigate regression tasks, and more specifically applications on the use of contrastive learning to improve regression hyperspectral data. In this paper, we propose a contrastive results on hyperspectral data, with application in hyperspectral learning framework for the regression tasks for hyperspectral unmixing and pollution estimation.