We explore feature space geometries induced by the 3-D Fourier scattering transform and deep neural network with extended attribute profiles on four standard hyperspectral images. We examine the distances and angles of class means, the variability of classes, and their low-dimensional structures. These statistics are compared to that of raw features, and our results provide insight into the vastly different properties of these two methods. We also explore a connection with the newly observed deep learning phenomenon of neural collapse.
@article{arxiv.2103.01303,
title = {Exploring the high dimensional geometry of HSI features},
author = {Wojciech Czaja and Ilya Kavalerov and Weilin Li},
journal= {arXiv preprint arXiv:2103.01303},
year = {2021}
}