English

Sequential Dimensionality Reduction for Extracting Localized Features

Computer Vision and Pattern Recognition 2016-10-07 v2 Machine Learning Numerical Analysis Numerical Analysis Machine Learning

Abstract

Linear dimensionality reduction techniques are powerful tools for image analysis as they allow the identification of important features in a data set. In particular, nonnegative matrix factorization (NMF) has become very popular as it is able to extract sparse, localized and easily interpretable features by imposing an additive combination of nonnegative basis elements. Nonnegative matrix underapproximation (NMU) is a closely related technique that has the advantage to identify features sequentially. In this paper, we propose a variant of NMU that is particularly well suited for image analysis as it incorporates the spatial information, that is, it takes into account the fact that neighboring pixels are more likely to be contained in the same features, and favors the extraction of localized features by looking for sparse basis elements. We show that our new approach competes favorably with comparable state-of-the-art techniques on synthetic, facial and hyperspectral image data sets.

Keywords

Cite

@article{arxiv.1505.06957,
  title  = {Sequential Dimensionality Reduction for Extracting Localized Features},
  author = {Gabriella Casalino and Nicolas Gillis},
  journal= {arXiv preprint arXiv:1505.06957},
  year   = {2016}
}

Comments

24 pages, 12 figures. New numerical experiments on synthetic data sets, discussion about the convergence

R2 v1 2026-06-22T09:41:32.798Z