Abstract
In this short article, we briefly retrospect the recent progress of spectral graph neural networks with manifold-learning-based feature extraction for hyperspectral image classification.