Progressive multi-scale multi-attention fusion for hyperspectral image classification

Abstract In recent years, due to the unique spatial-spectral characteristics of hyperspectral images, they have played a crucial role in many fields. The effective extraction of features using deep neural networks, followed by the design of efficient and high-precision network algorithm structures,...

Full description

Saved in:
Bibliographic Details
Main Authors: Hu Wang, Sixiang Quan, Jun Liu, Hai Xiao, Yingying Peng, Zhihui Wang, Huali Li
Format: Article
Language:English
Published: Nature Portfolio 2025-08-01
Series:Scientific Reports
Online Access:https://doi.org/10.1038/s41598-025-14844-w
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849235090418171904
author Hu Wang
Sixiang Quan
Jun Liu
Hai Xiao
Yingying Peng
Zhihui Wang
Huali Li
author_facet Hu Wang
Sixiang Quan
Jun Liu
Hai Xiao
Yingying Peng
Zhihui Wang
Huali Li
author_sort Hu Wang
collection DOAJ
description Abstract In recent years, due to the unique spatial-spectral characteristics of hyperspectral images, they have played a crucial role in many fields. The effective extraction of features using deep neural networks, followed by the design of efficient and high-precision network algorithm structures, has gradually become a research hotspot. Hyperspectral images are difficult to obtain and have limited samples. Although hyperspectral image classification methods based on convolutional neural networks (CNN) have noticeably improved performance, there are still certain shortcomings in the extraction of detailed and local features. Therefore, how to fully utilize spatial and spectral information in situations with limited samples has become a challenging problem. To address this issue, inspired by the PID controller, this paper proposes a Progressive Multi-Scale Multi-Attention Fusion (PMMF) network structure that simultaneously extracts features from the Proportional (P), Integral (I), and Derivative (D) branches. The complementary responsibilities of the three branches address the issue of feature loss in details and improve the network’s learning efficiency across feature maps of different scales. By cleverly extracting features from different branches multiple times, the fusion of multi-scale features is achieved, avoiding the limitations of single-scale feature representation. The proposed multi-attention fusion module applies the most suitable attention mechanism according to the representation form of each branch, fully extracting features from each branch, enriching the information contained in the feature maps, and greatly enhancing the classification accuracy of hyperspectral images.
format Article
id doaj-art-7aa0507b6d4b4778b8e39db8fbc2002f
institution Kabale University
issn 2045-2322
language English
publishDate 2025-08-01
publisher Nature Portfolio
record_format Article
series Scientific Reports
spelling doaj-art-7aa0507b6d4b4778b8e39db8fbc2002f2025-08-20T04:02:55ZengNature PortfolioScientific Reports2045-23222025-08-0115111910.1038/s41598-025-14844-wProgressive multi-scale multi-attention fusion for hyperspectral image classificationHu Wang0Sixiang Quan1Jun Liu2Hai Xiao3Yingying Peng4Zhihui Wang5Huali Li6School of Informatics, Hunan University of Chinese MedicineSecond Surveying and Mapping Institute of Hunan ProvinceSchool of Informatics, Hunan University of Chinese MedicineSecond Surveying and Mapping Institute of Hunan ProvinceSchool of Informatics, Hunan University of Chinese MedicineSchool of Informatics, Hunan University of Chinese MedicineCollege of Electrical and Information Engineering, Hunan UniversityAbstract In recent years, due to the unique spatial-spectral characteristics of hyperspectral images, they have played a crucial role in many fields. The effective extraction of features using deep neural networks, followed by the design of efficient and high-precision network algorithm structures, has gradually become a research hotspot. Hyperspectral images are difficult to obtain and have limited samples. Although hyperspectral image classification methods based on convolutional neural networks (CNN) have noticeably improved performance, there are still certain shortcomings in the extraction of detailed and local features. Therefore, how to fully utilize spatial and spectral information in situations with limited samples has become a challenging problem. To address this issue, inspired by the PID controller, this paper proposes a Progressive Multi-Scale Multi-Attention Fusion (PMMF) network structure that simultaneously extracts features from the Proportional (P), Integral (I), and Derivative (D) branches. The complementary responsibilities of the three branches address the issue of feature loss in details and improve the network’s learning efficiency across feature maps of different scales. By cleverly extracting features from different branches multiple times, the fusion of multi-scale features is achieved, avoiding the limitations of single-scale feature representation. The proposed multi-attention fusion module applies the most suitable attention mechanism according to the representation form of each branch, fully extracting features from each branch, enriching the information contained in the feature maps, and greatly enhancing the classification accuracy of hyperspectral images.https://doi.org/10.1038/s41598-025-14844-w
spellingShingle Hu Wang
Sixiang Quan
Jun Liu
Hai Xiao
Yingying Peng
Zhihui Wang
Huali Li
Progressive multi-scale multi-attention fusion for hyperspectral image classification
Scientific Reports
title Progressive multi-scale multi-attention fusion for hyperspectral image classification
title_full Progressive multi-scale multi-attention fusion for hyperspectral image classification
title_fullStr Progressive multi-scale multi-attention fusion for hyperspectral image classification
title_full_unstemmed Progressive multi-scale multi-attention fusion for hyperspectral image classification
title_short Progressive multi-scale multi-attention fusion for hyperspectral image classification
title_sort progressive multi scale multi attention fusion for hyperspectral image classification
url https://doi.org/10.1038/s41598-025-14844-w
work_keys_str_mv AT huwang progressivemultiscalemultiattentionfusionforhyperspectralimageclassification
AT sixiangquan progressivemultiscalemultiattentionfusionforhyperspectralimageclassification
AT junliu progressivemultiscalemultiattentionfusionforhyperspectralimageclassification
AT haixiao progressivemultiscalemultiattentionfusionforhyperspectralimageclassification
AT yingyingpeng progressivemultiscalemultiattentionfusionforhyperspectralimageclassification
AT zhihuiwang progressivemultiscalemultiattentionfusionforhyperspectralimageclassification
AT hualili progressivemultiscalemultiattentionfusionforhyperspectralimageclassification