Class-Aware Domain Adaptation for Coastal Land Cover Mapping Using Optical Remote Sensing Imagery

Coastal land cover mapping is a significant yet challenging pixel-level segmentation task. Domain shift between optical remote sensing imagery will give rise to remarkable performance degradation for deep supervised methods. Besides, the ground objects characterized with interclass variance and clas...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Jifa Chen, Gang Chen, Bo Fang, Jingjing Wang, Lizhe Wang
Formato:	article
Lenguaje:	EN
Publicado:	IEEE 2021
Materias:	Adversarial learning class balancing coastal land cover mapping (CLCM) domain adaptation entropy minimization (EM) Ocean engineering TC1501-1800 Geophysics. Cosmic physics QC801-809
Acceso en línea:	https://doaj.org/article/bda2a25762ef4368b993766f280ea435
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:bda2a25762ef4368b993766f280ea435
record_format	dspace
spelling	oai:doaj.org-article:bda2a25762ef4368b993766f280ea4352021-12-02T00:00:09ZClass-Aware Domain Adaptation for Coastal Land Cover Mapping Using Optical Remote Sensing Imagery2151-153510.1109/JSTARS.2021.3128527https://doaj.org/article/bda2a25762ef4368b993766f280ea4352021-01-01T00:00:00Zhttps://ieeexplore.ieee.org/document/9616474/https://doaj.org/toc/2151-1535Coastal land cover mapping is a significant yet challenging pixel-level segmentation task. Domain shift between optical remote sensing imagery will give rise to remarkable performance degradation for deep supervised methods. Besides, the ground objects characterized with interclass variance and class imbalance may further aggravate the adverse effect. Traditional adversary-based domain adaptation algorithms always leverage a binary discriminator to conduct global adaptation, ignoring the detailed class information. In this article, we develop a novel class-aware domain adaptation method to address these issues. Unlike the naive single one, we propose a joint local and global adversarial adaptation framework to separately execute class-specific and global domain alignment on feature and output spaces. For the former, the introduced classwise discriminator possesses different strategies to extract labels for both data domains. Meanwhile, we restore to entropy minimization to produce high-confident target prediction rather than using the early generated pseudo label with high confidence. Furthermore, we additionally adopt comprehensive reweighting on the supervised segmentation loss to track the class imbalance problem. This manner mainly comprises the sample-based median frequency balancing and the focal loss function for the minority and hard classes. We measure the proposed method on two typical coastal datasets and compare it with other state-of-the-art models. The experimental results confirm its excellent and competitive performance on cross-domain land cover mapping.Jifa ChenGang ChenBo FangJingjing WangLizhe WangIEEEarticleAdversarial learningclass balancingcoastal land cover mapping (CLCM)domain adaptationentropy minimization (EM)Ocean engineeringTC1501-1800Geophysics. Cosmic physicsQC801-809ENIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, Vol 14, Pp 11800-11813 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Adversarial learning class balancing coastal land cover mapping (CLCM) domain adaptation entropy minimization (EM) Ocean engineering TC1501-1800 Geophysics. Cosmic physics QC801-809
spellingShingle	Adversarial learning class balancing coastal land cover mapping (CLCM) domain adaptation entropy minimization (EM) Ocean engineering TC1501-1800 Geophysics. Cosmic physics QC801-809 Jifa Chen Gang Chen Bo Fang Jingjing Wang Lizhe Wang Class-Aware Domain Adaptation for Coastal Land Cover Mapping Using Optical Remote Sensing Imagery
description	Coastal land cover mapping is a significant yet challenging pixel-level segmentation task. Domain shift between optical remote sensing imagery will give rise to remarkable performance degradation for deep supervised methods. Besides, the ground objects characterized with interclass variance and class imbalance may further aggravate the adverse effect. Traditional adversary-based domain adaptation algorithms always leverage a binary discriminator to conduct global adaptation, ignoring the detailed class information. In this article, we develop a novel class-aware domain adaptation method to address these issues. Unlike the naive single one, we propose a joint local and global adversarial adaptation framework to separately execute class-specific and global domain alignment on feature and output spaces. For the former, the introduced classwise discriminator possesses different strategies to extract labels for both data domains. Meanwhile, we restore to entropy minimization to produce high-confident target prediction rather than using the early generated pseudo label with high confidence. Furthermore, we additionally adopt comprehensive reweighting on the supervised segmentation loss to track the class imbalance problem. This manner mainly comprises the sample-based median frequency balancing and the focal loss function for the minority and hard classes. We measure the proposed method on two typical coastal datasets and compare it with other state-of-the-art models. The experimental results confirm its excellent and competitive performance on cross-domain land cover mapping.
format	article
author	Jifa Chen Gang Chen Bo Fang Jingjing Wang Lizhe Wang
author_facet	Jifa Chen Gang Chen Bo Fang Jingjing Wang Lizhe Wang
author_sort	Jifa Chen
title	Class-Aware Domain Adaptation for Coastal Land Cover Mapping Using Optical Remote Sensing Imagery
title_short	Class-Aware Domain Adaptation for Coastal Land Cover Mapping Using Optical Remote Sensing Imagery
title_full	Class-Aware Domain Adaptation for Coastal Land Cover Mapping Using Optical Remote Sensing Imagery
title_fullStr	Class-Aware Domain Adaptation for Coastal Land Cover Mapping Using Optical Remote Sensing Imagery
title_full_unstemmed	Class-Aware Domain Adaptation for Coastal Land Cover Mapping Using Optical Remote Sensing Imagery
title_sort	class-aware domain adaptation for coastal land cover mapping using optical remote sensing imagery
publisher	IEEE
publishDate	2021
url	https://doaj.org/article/bda2a25762ef4368b993766f280ea435
work_keys_str_mv	AT jifachen classawaredomainadaptationforcoastallandcovermappingusingopticalremotesensingimagery AT gangchen classawaredomainadaptationforcoastallandcovermappingusingopticalremotesensingimagery AT bofang classawaredomainadaptationforcoastallandcovermappingusingopticalremotesensingimagery AT jingjingwang classawaredomainadaptationforcoastallandcovermappingusingopticalremotesensingimagery AT lizhewang classawaredomainadaptationforcoastallandcovermappingusingopticalremotesensingimagery
_version_	1718404011394072576

Class-Aware Domain Adaptation for Coastal Land Cover Mapping Using Optical Remote Sensing Imagery

Ejemplares similares