A Supervised Geometry-Aware Mapping Approach for Classification of Hyperspectral Images

TL;DR

This paper introduces a supervised, geometry-aware linear transformation method that improves hyperspectral image classification by enhancing class discrimination through intrinsic data structure preservation.

Contribution

It proposes a novel optimal geometry-aware transformation that leverages data geometry for better class separation in hyperspectral image classification.

Findings

Outperforms existing methods in classification accuracy

Effective in dimensionality reduction while maintaining discrimination

Validated on three real-world hyperspectral datasets

Abstract

The lack of proper class discrimination among the Hyperspectral (HS) data points poses a potential challenge in HS classification. To address this issue, this paper proposes an optimal geometry-aware transformation for enhancing the classification accuracy. The underlying idea of this method is to obtain a linear projection matrix by solving a nonlinear objective function based on the intrinsic geometrical structure of the data. The objective function is constructed to quantify the discrimination between the points from dissimilar classes on the projected data space. Then the obtained projection matrix is used to linearly map the data to more discriminative space. The effectiveness of the proposed transformation is illustrated with three benchmark real-world HS data sets. The experiments reveal that the classification and dimensionality reduction methods on the projected discriminative space outperform their counterpart in the original space.