Enhancing Binary Convolutional Neural Networks for Hyperspectral Image Classification

Hyperspectral remote sensing technology is swiftly evolving, prioritizing affordability, enhanced portability, seamless integration, sophisticated intelligence, and immediate processing capabilities. The leading model for classifying hyperspectral images, which relies on convolutional neural network...

Full description

Saved in:
Bibliographic Details
Main Authors: Xuebin Tang, Ke Zhang, Xiaolei Zhou, Lingbin Zeng, Shan Huang
Format: Article
Language:English
Published: MDPI AG 2024-11-01
Series:Remote Sensing
Subjects:
Online Access:https://www.mdpi.com/2072-4292/16/23/4398
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850259947175018496
author Xuebin Tang
Ke Zhang
Xiaolei Zhou
Lingbin Zeng
Shan Huang
author_facet Xuebin Tang
Ke Zhang
Xiaolei Zhou
Lingbin Zeng
Shan Huang
author_sort Xuebin Tang
collection DOAJ
description Hyperspectral remote sensing technology is swiftly evolving, prioritizing affordability, enhanced portability, seamless integration, sophisticated intelligence, and immediate processing capabilities. The leading model for classifying hyperspectral images, which relies on convolutional neural networks (CNNs), has proven to be highly effective when run on advanced computing platforms. Nonetheless, the high degree of parameterization inherent in CNN models necessitates considerable computational and storage resources, posing challenges to their deployment in processors with limited capacity like drones and satellites. This paper focuses on advancing lightweight models for hyperspectral image classification and introduces EBCNN, a novel binary convolutional neural network. EBCNN is designed to effectively regulate backpropagation gradients and minimize gradient discrepancies to optimize BNN performance. EBCNN incorporates an adaptive gradient scaling module that utilizes a multi-scale pyramid squeeze attention (PSA) mechanism during the training phase, which can adjust training gradients flexibly and efficiently. Additionally, to address suboptimal training issues, EBCNN employs a dynamic curriculum learning strategy underpinned by a confidence-aware loss function, Superloss, enabling progressive binarization and enhancing its classification effectiveness. Extensive experimental evaluations conducted on five esteemed public datasets confirm the effectiveness of EBCNN. These analyses highlight a significant enhancement in the classification accuracy of hyperspectral images, achieved without incurring additional memory or computational overheads during the inference process.
format Article
id doaj-art-8e3f9fa10f7b411fb4a86d68c9eabe2a
institution OA Journals
issn 2072-4292
language English
publishDate 2024-11-01
publisher MDPI AG
record_format Article
series Remote Sensing
spelling doaj-art-8e3f9fa10f7b411fb4a86d68c9eabe2a2025-08-20T01:55:45ZengMDPI AGRemote Sensing2072-42922024-11-011623439810.3390/rs16234398Enhancing Binary Convolutional Neural Networks for Hyperspectral Image ClassificationXuebin Tang0Ke Zhang1Xiaolei Zhou2Lingbin Zeng3Shan Huang4National University of Defense Technology, ChinaNational University of Defense Technology, ChinaNational University of Defense Technology, ChinaNational University of Defense Technology, ChinaNational University of Defense Technology, ChinaHyperspectral remote sensing technology is swiftly evolving, prioritizing affordability, enhanced portability, seamless integration, sophisticated intelligence, and immediate processing capabilities. The leading model for classifying hyperspectral images, which relies on convolutional neural networks (CNNs), has proven to be highly effective when run on advanced computing platforms. Nonetheless, the high degree of parameterization inherent in CNN models necessitates considerable computational and storage resources, posing challenges to their deployment in processors with limited capacity like drones and satellites. This paper focuses on advancing lightweight models for hyperspectral image classification and introduces EBCNN, a novel binary convolutional neural network. EBCNN is designed to effectively regulate backpropagation gradients and minimize gradient discrepancies to optimize BNN performance. EBCNN incorporates an adaptive gradient scaling module that utilizes a multi-scale pyramid squeeze attention (PSA) mechanism during the training phase, which can adjust training gradients flexibly and efficiently. Additionally, to address suboptimal training issues, EBCNN employs a dynamic curriculum learning strategy underpinned by a confidence-aware loss function, Superloss, enabling progressive binarization and enhancing its classification effectiveness. Extensive experimental evaluations conducted on five esteemed public datasets confirm the effectiveness of EBCNN. These analyses highlight a significant enhancement in the classification accuracy of hyperspectral images, achieved without incurring additional memory or computational overheads during the inference process.https://www.mdpi.com/2072-4292/16/23/4398remote sensinghyperspectral image classificationbinary neural networkconvolutional neural networkdynamic curriculum learningpyramid squeeze attention
spellingShingle Xuebin Tang
Ke Zhang
Xiaolei Zhou
Lingbin Zeng
Shan Huang
Enhancing Binary Convolutional Neural Networks for Hyperspectral Image Classification
Remote Sensing
remote sensing
hyperspectral image classification
binary neural network
convolutional neural network
dynamic curriculum learning
pyramid squeeze attention
title Enhancing Binary Convolutional Neural Networks for Hyperspectral Image Classification
title_full Enhancing Binary Convolutional Neural Networks for Hyperspectral Image Classification
title_fullStr Enhancing Binary Convolutional Neural Networks for Hyperspectral Image Classification
title_full_unstemmed Enhancing Binary Convolutional Neural Networks for Hyperspectral Image Classification
title_short Enhancing Binary Convolutional Neural Networks for Hyperspectral Image Classification
title_sort enhancing binary convolutional neural networks for hyperspectral image classification
topic remote sensing
hyperspectral image classification
binary neural network
convolutional neural network
dynamic curriculum learning
pyramid squeeze attention
url https://www.mdpi.com/2072-4292/16/23/4398
work_keys_str_mv AT xuebintang enhancingbinaryconvolutionalneuralnetworksforhyperspectralimageclassification
AT kezhang enhancingbinaryconvolutionalneuralnetworksforhyperspectralimageclassification
AT xiaoleizhou enhancingbinaryconvolutionalneuralnetworksforhyperspectralimageclassification
AT lingbinzeng enhancingbinaryconvolutionalneuralnetworksforhyperspectralimageclassification
AT shanhuang enhancingbinaryconvolutionalneuralnetworksforhyperspectralimageclassification