LEt-SNE: A Hybrid Approach To Data Embedding and Visualization of Hyperspectral Imagery

TL;DR

LEt-SNE is a hybrid dimensionality reduction technique combining t-SNE and Laplacian Eigenmaps, optimized for hyperspectral imagery to improve visualization and clustering while addressing high-dimensional challenges.

Contribution

The paper introduces LEt-SNE, a novel hybrid algorithm with a new Compression Factor term, specifically designed for hyperspectral data to enhance manifold visualization and clustering.

Findings

LEt-SNE performs competitively with state-of-the-art methods on hyperspectral datasets.

The Compression Factor effectively mitigates the curse of dimensionality.

The method is suitable for both labeled and unlabeled data visualization.

Abstract

Hyperspectral Imagery (and Remote Sensing in general) captured from UAVs or satellites are highly voluminous in nature due to the large spatial extent and wavelengths captured by them. Since analyzing these images requires a huge amount of computational time and power, various dimensionality reduction techniques have been used for feature reduction. Some popular techniques among these falter when applied to Hyperspectral Imagery due to the famed curse of dimensionality. In this paper, we propose a novel approach, LEt-SNE, which combines graph based algorithms like t-SNE and Laplacian Eigenmaps into a model parameterized by a shallow feed forward network. We introduce a new term, Compression Factor, that enables our method to combat the curse of dimensionality. The proposed algorithm is suitable for manifold visualization and sample clustering with labelled or unlabelled data. We demonstrate that our method is competitive with current state-of-the-art methods on hyperspectral remote sensing datasets in public domain.