Semi-supervised Learning with Graphs: Covariance Based Superpixels For Hyperspectral Image Classification

Philip Sellars , Angelica I. Aviles-Rivero (1) , Nicolas Papadakis (2) , David Coomes (3) , Anita Faul , Carola-Bibiane Schönlieb (1)

1 DAMTP - Department of Applied Mathematics and Theoretical Physics
2 IMB - Institut de Mathématiques de Bordeaux
3 Plant Sciences, University of Cambridge

In this paper, we present a graph-based semi-supervised framework for hyperspectral image classification. We first introduce a novel superpixel algorithm based on the spectral covariance matrix representation of pixels to provide a better representation of our data. We then construct a superpixel graph, based on carefully considered feature vectors, before performing classification. We demonstrate, through a set of experimental results using two benchmarking datasets, that our approach outperforms three state-of-the-art classification frameworks, especially when an extremely small amount of labelled data is used.

Mots clés: Index Terms-Hyperspectral Imaging; Superpixels; Covariance; Graphs; Semi-Supervised Learning; Classification

Domaines: Traitement du signal et de l'image [eess.SP]

Philip Sellars, Angelica I. Aviles-Rivero, Nicolas Papadakis, David Coomes, Anita Faul, et al.. Semi-supervised Learning with Graphs: Covariance Based Superpixels For Hyperspectral Image Classification. IEEE Geoscience and Remote Sensing Symposium (IGARSS'19), Jul 2019, Yokohama, Japan. pp.592-595. ⟨hal-02057730⟩