Local-Descriptors-Based Rectification Network for Few-Shot Remote Sensing Scene Classification

Few-shot remote sensing scene classification has become a study that has attracted widespread attention and aims to identify new scene classes through one or a few labeled scene images. Nevertheless, due to the existence of unrelated complex background in scene images, local descriptors (LDs) that o...

Full description

Saved in:
Bibliographic Details
Main Authors: Anyong Qin, Bin Luo, Qiang Li, Cuiming Zou, Yu Zhao, Tiecheng Song, Chenqiang Gao
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10925636/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850199639793336320
author Anyong Qin
Bin Luo
Qiang Li
Cuiming Zou
Yu Zhao
Tiecheng Song
Chenqiang Gao
author_facet Anyong Qin
Bin Luo
Qiang Li
Cuiming Zou
Yu Zhao
Tiecheng Song
Chenqiang Gao
author_sort Anyong Qin
collection DOAJ
description Few-shot remote sensing scene classification has become a study that has attracted widespread attention and aims to identify new scene classes through one or a few labeled scene images. Nevertheless, due to the existence of unrelated complex background in scene images, local descriptors (LDs) that offer a more efficient representation than image-level features, will carry semantic information unrelated to the real semantics of the scene images. Concurrently, these irrelevant background LDs are also causing a large distribution bias in support and query sets, which leads to the problem of inaccurate feature representation of scene images. To address the aforementioned problems, in this article, we introduce an LD-based rectification network called LDRNet. Within this network, we first design an LD semantic rectification module. It performs semantic rectification on LDs that are unrelated to scene image semantics by obtaining a descriptor-level global-aware semantic representation. Second, we introduce a cross-set bias rectification module. It rectifies the query set by obtaining the offset between two sets (query and support) from a more detailed LD perspective. This operation can shorten the distance among the two sets (query and support), thereby obtaining a more accurate representation of scene image features. Furthermore, we employ an LD-based contrastive loss function to guarantee that the rectified LD semantics are consistent with the corresponding scene image. The comparative experimental result indicates that our LDRNet achieves state-of-the-art performance on three commonly used public datasets.
format Article
id doaj-art-292674282db4479dbc5357d5fd1c7e73
institution OA Journals
issn 1939-1404
2151-1535
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
spelling doaj-art-292674282db4479dbc5357d5fd1c7e732025-08-20T02:12:34ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-01189566958110.1109/JSTARS.2025.355159910925636Local-Descriptors-Based Rectification Network for Few-Shot Remote Sensing Scene ClassificationAnyong Qin0https://orcid.org/0000-0002-2538-822XBin Luo1https://orcid.org/0009-0007-7941-2479Qiang Li2Cuiming Zou3https://orcid.org/0000-0002-2283-9048Yu Zhao4https://orcid.org/0000-0001-6593-542XTiecheng Song5https://orcid.org/0000-0003-1264-2812Chenqiang Gao6https://orcid.org/0000-0003-4174-4148School of Communications and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing, ChinaSchool of Communications and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing, ChinaSchool of Communications and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing, ChinaCollege of Informatics, Huazhong Agricultural University, Wuhan, ChinaSchool of Communications and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing, ChinaSchool of Communications and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing, ChinaSchool of Intelligent Systems Engineering, Sun Yat-Sen University, Shenzhen, ChinaFew-shot remote sensing scene classification has become a study that has attracted widespread attention and aims to identify new scene classes through one or a few labeled scene images. Nevertheless, due to the existence of unrelated complex background in scene images, local descriptors (LDs) that offer a more efficient representation than image-level features, will carry semantic information unrelated to the real semantics of the scene images. Concurrently, these irrelevant background LDs are also causing a large distribution bias in support and query sets, which leads to the problem of inaccurate feature representation of scene images. To address the aforementioned problems, in this article, we introduce an LD-based rectification network called LDRNet. Within this network, we first design an LD semantic rectification module. It performs semantic rectification on LDs that are unrelated to scene image semantics by obtaining a descriptor-level global-aware semantic representation. Second, we introduce a cross-set bias rectification module. It rectifies the query set by obtaining the offset between two sets (query and support) from a more detailed LD perspective. This operation can shorten the distance among the two sets (query and support), thereby obtaining a more accurate representation of scene image features. Furthermore, we employ an LD-based contrastive loss function to guarantee that the rectified LD semantics are consistent with the corresponding scene image. The comparative experimental result indicates that our LDRNet achieves state-of-the-art performance on three commonly used public datasets.https://ieeexplore.ieee.org/document/10925636/Bias rectificationfew-shot learning (FSL)local descriptor (LD)remote sensing scene classification (RSSC)semantics rectification
spellingShingle Anyong Qin
Bin Luo
Qiang Li
Cuiming Zou
Yu Zhao
Tiecheng Song
Chenqiang Gao
Local-Descriptors-Based Rectification Network for Few-Shot Remote Sensing Scene Classification
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Bias rectification
few-shot learning (FSL)
local descriptor (LD)
remote sensing scene classification (RSSC)
semantics rectification
title Local-Descriptors-Based Rectification Network for Few-Shot Remote Sensing Scene Classification
title_full Local-Descriptors-Based Rectification Network for Few-Shot Remote Sensing Scene Classification
title_fullStr Local-Descriptors-Based Rectification Network for Few-Shot Remote Sensing Scene Classification
title_full_unstemmed Local-Descriptors-Based Rectification Network for Few-Shot Remote Sensing Scene Classification
title_short Local-Descriptors-Based Rectification Network for Few-Shot Remote Sensing Scene Classification
title_sort local descriptors based rectification network for few shot remote sensing scene classification
topic Bias rectification
few-shot learning (FSL)
local descriptor (LD)
remote sensing scene classification (RSSC)
semantics rectification
url https://ieeexplore.ieee.org/document/10925636/
work_keys_str_mv AT anyongqin localdescriptorsbasedrectificationnetworkforfewshotremotesensingsceneclassification
AT binluo localdescriptorsbasedrectificationnetworkforfewshotremotesensingsceneclassification
AT qiangli localdescriptorsbasedrectificationnetworkforfewshotremotesensingsceneclassification
AT cuimingzou localdescriptorsbasedrectificationnetworkforfewshotremotesensingsceneclassification
AT yuzhao localdescriptorsbasedrectificationnetworkforfewshotremotesensingsceneclassification
AT tiechengsong localdescriptorsbasedrectificationnetworkforfewshotremotesensingsceneclassification
AT chenqianggao localdescriptorsbasedrectificationnetworkforfewshotremotesensingsceneclassification