Linear Data Compression of Hyperspectral Images
Abstract
The aim of the paper is to analyse hyperspectral images using tensor principal component analysis of multi-way data sets. The mathematical and computational backgrounds of pattern recognition are the geometries in Hilbert space for functional analysis and applied linear algebra for numerical analysis, respectively. Because of high-resolution sampling in the colour channels, images observed by a hyperspectral camera system are expressed by three-mode tensors. The Tucker-3 decomposition of a three-mode tensor is used in behaviourmetric science and psychology for the extraction of relations among three entries as an extension of the usual principal component analysis for statistical analysis. Hyperspectral images express spectral information of two-dimensional images on the imaging plane. Therefore, for statistical analysis, we adopt the Tucker-3 decomposition. The Tucker-3 decomposition of hyperspectral images extracts statistically dominant information from hyperspectral images. Tensor principal component analysis allows us to extract dominant light-channel information from hyperspectral images.
Cite
Text
Tanji et al. "Linear Data Compression of Hyperspectral Images." IEEE/CVF International Conference on Computer Vision Workshops, 2017. doi:10.1109/ICCVW.2017.354Markdown
[Tanji et al. "Linear Data Compression of Hyperspectral Images." IEEE/CVF International Conference on Computer Vision Workshops, 2017.](https://mlanthology.org/iccvw/2017/tanji2017iccvw-linear/) doi:10.1109/ICCVW.2017.354BibTeX
@inproceedings{tanji2017iccvw-linear,
title = {{Linear Data Compression of Hyperspectral Images}},
author = {Tanji, Kaori and Imiya, Atsushi and Itoh, Hayato and Kuze, Hiroaki and Manago, Naohiro},
booktitle = {IEEE/CVF International Conference on Computer Vision Workshops},
year = {2017},
pages = {3001-3007},
doi = {10.1109/ICCVW.2017.354},
url = {https://mlanthology.org/iccvw/2017/tanji2017iccvw-linear/}
}