Human-Annotated Label Noise and Their Impact on ConvNets for Remote Sensing Image Scene Classification

Human-labeled training datasets are essential for convolutional neural networks (ConvNets) in satellite image scene classification. Annotation errors are unavoidable due to the complexity of satellite images. However, the distribution of real-world human-annotated label noises on satellite images an...

Full description

Saved in:
Bibliographic Details
Main Authors: Longkang Peng, Tao Wei, Xuehong Chen, Xiaobei Chen, Rui Sun, Luoma Wan, Jin Chen, Xiaolin Zhu
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/10757363/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850106317635584000
author Longkang Peng
Tao Wei
Xuehong Chen
Xiaobei Chen
Rui Sun
Luoma Wan
Jin Chen
Xiaolin Zhu
author_facet Longkang Peng
Tao Wei
Xuehong Chen
Xiaobei Chen
Rui Sun
Luoma Wan
Jin Chen
Xiaolin Zhu
author_sort Longkang Peng
collection DOAJ
description Human-labeled training datasets are essential for convolutional neural networks (ConvNets) in satellite image scene classification. Annotation errors are unavoidable due to the complexity of satellite images. However, the distribution of real-world human-annotated label noises on satellite images and their impact on ConvNets have not been investigated. To fill this research gap, this article, for the first time, collected real-world labels from 32 participants and explored how their annotated label noise affects three representative ConvNets (VGG16, GoogleNet, and ResNet-50) for remote sensing image scene classification. We found that 1) human-annotated label noise exhibits significant class and instance dependence; 2) an additional 1% of human-annotated label noise in training data leads to a 0.5% reduction in the overall accuracy of ConvNets classification; and 3) the error pattern of ConvNet predictions was strongly correlated with that of participant's labels. To uncover the mechanism underlying the impact of human labeling errors on ConvNets, we compared it with three types of simulated label noise: uniform noise, class-dependent noise, and instance-dependent noise. Our results show that the impact of human-annotated label noise on ConvNets significantly differs from all three types of simulated label noise, while both class dependence and instance dependence contribute to the impact of human-annotated label noise on ConvNets. Additionally, the label noise estimation algorithm (confident learning) cannot fully identify label noise. These observations necessitate a reevaluation of the handling of noisy labels, and we anticipate that our real-world label noise dataset would facilitate the future development and assessment of label-noise learning algorithms.
format Article
id doaj-art-45cae411c52b4aeda5b1ff5cada15de8
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-45cae411c52b4aeda5b1ff5cada15de82025-08-20T02:38:51ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-01181500151410.1109/JSTARS.2024.350246110757363Human-Annotated Label Noise and Their Impact on ConvNets for Remote Sensing Image Scene ClassificationLongkang Peng0https://orcid.org/0009-0003-0015-0199Tao Wei1https://orcid.org/0000-0001-8094-4906Xuehong Chen2https://orcid.org/0000-0001-7223-8649Xiaobei Chen3Rui Sun4Luoma Wan5https://orcid.org/0000-0002-8202-510XJin Chen6https://orcid.org/0000-0002-6497-4141Xiaolin Zhu7https://orcid.org/0000-0001-6967-786XSchool of Psychology, Shenzhen University, Shenzhen, ChinaSchool of Psychology, Shenzhen University, Shenzhen, ChinaState Key Laboratory of Remote Sensing Science, Faculty of Geographical Science, Beijing Normal University, Beijing, ChinaSchool of Psychology, Shenzhen University, Shenzhen, ChinaSchool of Psychology, Shenzhen University, Shenzhen, ChinaDepartment of Land Surveying and Geo-Informatics and the Otto Poon Charitable Foundation Smart Cities Research Institute, The Hong Kong Polytechnic University, Hong KongState Key Laboratory of Remote Sensing Science, Faculty of Geographical Science, Beijing Normal University, Beijing, ChinaDepartment of Land Surveying and Geo-Informatics and the Otto Poon Charitable Foundation Smart Cities Research Institute, The Hong Kong Polytechnic University, Hong KongHuman-labeled training datasets are essential for convolutional neural networks (ConvNets) in satellite image scene classification. Annotation errors are unavoidable due to the complexity of satellite images. However, the distribution of real-world human-annotated label noises on satellite images and their impact on ConvNets have not been investigated. To fill this research gap, this article, for the first time, collected real-world labels from 32 participants and explored how their annotated label noise affects three representative ConvNets (VGG16, GoogleNet, and ResNet-50) for remote sensing image scene classification. We found that 1) human-annotated label noise exhibits significant class and instance dependence; 2) an additional 1% of human-annotated label noise in training data leads to a 0.5% reduction in the overall accuracy of ConvNets classification; and 3) the error pattern of ConvNet predictions was strongly correlated with that of participant's labels. To uncover the mechanism underlying the impact of human labeling errors on ConvNets, we compared it with three types of simulated label noise: uniform noise, class-dependent noise, and instance-dependent noise. Our results show that the impact of human-annotated label noise on ConvNets significantly differs from all three types of simulated label noise, while both class dependence and instance dependence contribute to the impact of human-annotated label noise on ConvNets. Additionally, the label noise estimation algorithm (confident learning) cannot fully identify label noise. These observations necessitate a reevaluation of the handling of noisy labels, and we anticipate that our real-world label noise dataset would facilitate the future development and assessment of label-noise learning algorithms.https://ieeexplore.ieee.org/document/10757363/Convolutional neural network (ConvNet)human-annotated label noiselabel noiseremote sensingscene classification
spellingShingle Longkang Peng
Tao Wei
Xuehong Chen
Xiaobei Chen
Rui Sun
Luoma Wan
Jin Chen
Xiaolin Zhu
Human-Annotated Label Noise and Their Impact on ConvNets for Remote Sensing Image Scene Classification
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Convolutional neural network (ConvNet)
human-annotated label noise
label noise
remote sensing
scene classification
title Human-Annotated Label Noise and Their Impact on ConvNets for Remote Sensing Image Scene Classification
title_full Human-Annotated Label Noise and Their Impact on ConvNets for Remote Sensing Image Scene Classification
title_fullStr Human-Annotated Label Noise and Their Impact on ConvNets for Remote Sensing Image Scene Classification
title_full_unstemmed Human-Annotated Label Noise and Their Impact on ConvNets for Remote Sensing Image Scene Classification
title_short Human-Annotated Label Noise and Their Impact on ConvNets for Remote Sensing Image Scene Classification
title_sort human annotated label noise and their impact on convnets for remote sensing image scene classification
topic Convolutional neural network (ConvNet)
human-annotated label noise
label noise
remote sensing
scene classification
url https://ieeexplore.ieee.org/document/10757363/
work_keys_str_mv AT longkangpeng humanannotatedlabelnoiseandtheirimpactonconvnetsforremotesensingimagesceneclassification
AT taowei humanannotatedlabelnoiseandtheirimpactonconvnetsforremotesensingimagesceneclassification
AT xuehongchen humanannotatedlabelnoiseandtheirimpactonconvnetsforremotesensingimagesceneclassification
AT xiaobeichen humanannotatedlabelnoiseandtheirimpactonconvnetsforremotesensingimagesceneclassification
AT ruisun humanannotatedlabelnoiseandtheirimpactonconvnetsforremotesensingimagesceneclassification
AT luomawan humanannotatedlabelnoiseandtheirimpactonconvnetsforremotesensingimagesceneclassification
AT jinchen humanannotatedlabelnoiseandtheirimpactonconvnetsforremotesensingimagesceneclassification
AT xiaolinzhu humanannotatedlabelnoiseandtheirimpactonconvnetsforremotesensingimagesceneclassification