Linear vs Nonlinear Extreme Learning Machine for Spectral-Spatial Classification of Hyperspectral Image

TL;DR

This paper compares linear and nonlinear extreme learning machines for hyperspectral image classification, proposing a spectral-spatial framework that enhances accuracy by integrating belief propagation, with linear ELM outperforming nonlinear variants.

Contribution

It introduces a spectral-spatial classification framework combining linear ELM with loopy belief propagation, demonstrating the superiority of linear ELM for hyperspectral data.

Findings

Linear ELM outperforms nonlinear ELM in spectral-spatial classification.

The proposed method maintains fast speed while significantly improving accuracy.

Experimental results on Indian Pines and Pavia datasets validate the approach.

Abstract

As a new machine learning approach, extreme learning machine (ELM) has received wide attentions due to its good performances. However, when directly applied to the hyperspectral image (HSI) classification, the recognition rate is too low. This is because ELM does not use the spatial information which is very important for HSI classification. In view of this, this paper proposes a new framework for spectral-spatial classification of HSI by combining ELM with loopy belief propagation (LBP). The original ELM is linear, and the nonlinear ELMs (or Kernel ELMs) are the improvement of linear ELM (LELM). However, based on lots of experiments and analysis, we found out that the LELM is a better choice than nonlinear ELM for spectral-spatial classification of HSI. Furthermore, we exploit the marginal probability distribution that uses the whole information in the HSI and learn such distribution using the LBP. The proposed method not only maintain the fast speed of ELM, but also greatly improves the accuracy of classification. The experimental results in the well-known HSI data sets, Indian Pines and Pavia University, demonstrate the good performances of the proposed method.