Feature Fusion via Dual-resolution Compressive Measurement Matrix Analysis For Spectral Image Classification

TL;DR

This paper introduces a novel feature fusion method from dual-resolution compressive measurements for spectral image classification, improving accuracy by modeling measurement degradation and incorporating regularization techniques.

Contribution

It proposes a new fusion approach formulated as an inverse problem, utilizing coded aperture information and regularization to enhance spectral image classification from compressive data.

Findings

Outperforms existing methods in spectral image classification accuracy

Effectively models measurement degradation using coded aperture patterns

Incorporates sparsity and TV regularization for improved pixel correlation handling

Abstract

In the compressive spectral imaging (CSI) framework, different architectures have been proposed to recover high-resolution spectral images from compressive measurements. Since CSI architectures compactly capture the relevant information of the spectral image, various methods that extract classification features from compressive samples have been recently proposed. However, these techniques require a feature extraction procedure that reorders measurements using the information embedded in the coded aperture patterns. In this paper, a method that fuses features directly from dual-resolution compressive measurements is proposed for spectral image classification. More precisely, the fusion method is formulated as an inverse problem that estimates high-spatial-resolution and low-dimensional feature bands from compressive measurements. To this end, the decimation matrices that describe the compressive measurements as degraded versions of the fused features are mathematically modeled using the information embedded in the coded aperture patterns. Furthermore, we include both a sparsity-promoting and a total-variation (TV) regularization terms to the fusion problem in order to consider the correlations between neighbor pixels, and therefore, improve the accuracy of pixel-based classifiers. To solve the fusion problem, we describe an algorithm based on the accelerated variant of the alternating direction method of multipliers (accelerated-ADMM). Additionally, a classification approach that includes the developed fusion method and a multilayer neural network is introduced. Finally, the proposed approach is evaluated on three remote sensing spectral images and a set of compressive measurements captured in the laboratory. Extensive simulations show that the proposed classification approach outperforms other approaches under various performance metrics.