Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

SRL-SOA: Self-Representation Learning with Sparse 1D-Operational Autoencoder for Hyperspectral Image Band Selection

Mete Ahishali, Serkan Kiranyaz, Iftikhar Ahmad, Moncef Gabbouj

TL;DR

The proposed SLR-SOA approach introduces a novel autoencoder model, SOA, that is designed to learn a representation domain where the data are sparsely represented, and outperforms the competing methods over two HSI data including Indian Pines and Salinas-A considering the achieved land cover classification accuracies.

Abstract

The band selection in the hyperspectral image (HSI) data processing is an important task considering its effect on the computational complexity and accuracy. In this work, we propose a novel framework for the band selection problem: Self-Representation Learning (SRL) with Sparse 1D-Operational Autoencoder (SOA). The proposed SLR-SOA approach introduces a novel autoencoder model, SOA, that is designed to learn a representation domain where the data are sparsely represented. Moreover, the network composes of 1D-operational layers with the non-linear neuron model. Hence, the learning capability of neurons (filters) is greatly improved with shallow architectures. Using compact architectures is especially crucial in autoencoders as they tend to overfit easily because of their identity mapping objective. Overall, we show that the proposed SRL-SOA band selection approach outperforms the competing methods over two HSI data including Indian Pines and Salinas-A considering the achieved land cover classification accuracies. The software implementation of the SRL-SOA approach is shared publicly at https://github.com/meteahishali/SRL-SOA.

SRL-SOA: Self-Representation Learning with Sparse 1D-Operational Autoencoder for Hyperspectral Image Band Selection

TL;DR

The proposed SLR-SOA approach introduces a novel autoencoder model, SOA, that is designed to learn a representation domain where the data are sparsely represented, and outperforms the competing methods over two HSI data including Indian Pines and Salinas-A considering the achieved land cover classification accuracies.

Abstract

The band selection in the hyperspectral image (HSI) data processing is an important task considering its effect on the computational complexity and accuracy. In this work, we propose a novel framework for the band selection problem: Self-Representation Learning (SRL) with Sparse 1D-Operational Autoencoder (SOA). The proposed SLR-SOA approach introduces a novel autoencoder model, SOA, that is designed to learn a representation domain where the data are sparsely represented. Moreover, the network composes of 1D-operational layers with the non-linear neuron model. Hence, the learning capability of neurons (filters) is greatly improved with shallow architectures. Using compact architectures is especially crucial in autoencoders as they tend to overfit easily because of their identity mapping objective. Overall, we show that the proposed SRL-SOA band selection approach outperforms the competing methods over two HSI data including Indian Pines and Salinas-A considering the achieved land cover classification accuracies. The software implementation of the SRL-SOA approach is shared publicly at https://github.com/meteahishali/SRL-SOA.
Paper Structure (9 sections, 7 equations, 2 figures, 1 table, 1 algorithm)

This paper contains 9 sections, 7 equations, 2 figures, 1 table, 1 algorithm.

Table of Contents

  1. Introduction
  2. Proposed Methodology
  3. Sparse Self-Representation
  4. 1D-Operational Layers
  5. SRL-SOA: Self-Representation Learning with Sparse 1D-Operational Autoencoder
  6. Experimental Evaluation
  7. Datasets and Experimental Setup
  8. Results
  9. Conclusion

Figures (2)

  • Figure 1: The proposed SRL-SOA framework with the 1D-operational layer where $\mathbf{X}_s$ is the batch sampled HSI cube and $\mathbf{A}$ is the learned representation matrix. The batch size is set to $m=1$ for illustration purposes. The network filters operate as expressed in \ref{['eq:operation']} and it is trained with the loss function given in \ref{['cost']}.
  • Figure 2: The classification results versus the selected number of bands by the proposed approach ($\text{SRL-SOA}_Q$ with $Q=1$, $3$, and $5$) and different band selection methods on the Indian Pines dataset in (a - c) and Salinas-A dataset in (d - f).