Rotation-Invariant and Relation-Aware Cross-Domain Adaptation Object Detection Network for Optical Remote Sensing Images

In recent years, object detection has shown excellent results on a large number of annotated data, but when there is a discrepancy between the annotated data and the real test data, the performance of the trained object detection model is often degraded when it is directly transferred to the real te...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Ying Chen, Qi Liu, Teng Wang, Bin Wang, Xiaoliang Meng
Formato:	article
Lenguaje:	EN
Publicado:	MDPI AG 2021
Materias:	object detection unsupervised domain adaptation remote sensing images rotation invariance graph convolutional neural network (GCN) Science Q
Acceso en línea:	https://doaj.org/article/ed656ef591a44896bcc2c5286e5cf6fc
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:ed656ef591a44896bcc2c5286e5cf6fc
record_format	dspace
spelling	oai:doaj.org-article:ed656ef591a44896bcc2c5286e5cf6fc2021-11-11T18:55:06ZRotation-Invariant and Relation-Aware Cross-Domain Adaptation Object Detection Network for Optical Remote Sensing Images10.3390/rs132143862072-4292https://doaj.org/article/ed656ef591a44896bcc2c5286e5cf6fc2021-10-01T00:00:00Zhttps://www.mdpi.com/2072-4292/13/21/4386https://doaj.org/toc/2072-4292In recent years, object detection has shown excellent results on a large number of annotated data, but when there is a discrepancy between the annotated data and the real test data, the performance of the trained object detection model is often degraded when it is directly transferred to the real test dataset. Compared with natural images, remote sensing images have great differences in appearance and quality. Traditional methods need to re-label all image data before interpretation, which will consume a lot of manpower and time. Therefore, it is of practical significance to study the Cross-Domain Adaptation Object Detection (CDAOD) of remote sensing images. To solve the above problems, our paper proposes a Rotation-Invariant and Relation-Aware (RIRA) CDAOD network. We trained the network at the image-level and the prototype-level based on a relation aware graph to align the feature distribution and added the rotation-invariant regularizer to deal with the rotation diversity. The Faster R-CNN network was adopted as the backbone framework of the network. We conducted experiments on two typical remote sensing building detection datasets, and set three domain adaptation scenarios: WHU 2012 → WHU 2016, Inria (Chicago) → Inria (Austin), and WHU 2012 → Inria (Austin). The results show that our method can effectively improve the detection effect in the target domain, and outperform competing methods by obtaining optimal results in all three scenarios.Ying ChenQi LiuTeng WangBin WangXiaoliang MengMDPI AGarticleobject detectionunsupervised domain adaptationremote sensing imagesrotation invariancegraph convolutional neural network (GCN)ScienceQENRemote Sensing, Vol 13, Iss 4386, p 4386 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	object detection unsupervised domain adaptation remote sensing images rotation invariance graph convolutional neural network (GCN) Science Q
spellingShingle	object detection unsupervised domain adaptation remote sensing images rotation invariance graph convolutional neural network (GCN) Science Q Ying Chen Qi Liu Teng Wang Bin Wang Xiaoliang Meng Rotation-Invariant and Relation-Aware Cross-Domain Adaptation Object Detection Network for Optical Remote Sensing Images
description	In recent years, object detection has shown excellent results on a large number of annotated data, but when there is a discrepancy between the annotated data and the real test data, the performance of the trained object detection model is often degraded when it is directly transferred to the real test dataset. Compared with natural images, remote sensing images have great differences in appearance and quality. Traditional methods need to re-label all image data before interpretation, which will consume a lot of manpower and time. Therefore, it is of practical significance to study the Cross-Domain Adaptation Object Detection (CDAOD) of remote sensing images. To solve the above problems, our paper proposes a Rotation-Invariant and Relation-Aware (RIRA) CDAOD network. We trained the network at the image-level and the prototype-level based on a relation aware graph to align the feature distribution and added the rotation-invariant regularizer to deal with the rotation diversity. The Faster R-CNN network was adopted as the backbone framework of the network. We conducted experiments on two typical remote sensing building detection datasets, and set three domain adaptation scenarios: WHU 2012 → WHU 2016, Inria (Chicago) → Inria (Austin), and WHU 2012 → Inria (Austin). The results show that our method can effectively improve the detection effect in the target domain, and outperform competing methods by obtaining optimal results in all three scenarios.
format	article
author	Ying Chen Qi Liu Teng Wang Bin Wang Xiaoliang Meng
author_facet	Ying Chen Qi Liu Teng Wang Bin Wang Xiaoliang Meng
author_sort	Ying Chen
title	Rotation-Invariant and Relation-Aware Cross-Domain Adaptation Object Detection Network for Optical Remote Sensing Images
title_short	Rotation-Invariant and Relation-Aware Cross-Domain Adaptation Object Detection Network for Optical Remote Sensing Images
title_full	Rotation-Invariant and Relation-Aware Cross-Domain Adaptation Object Detection Network for Optical Remote Sensing Images
title_fullStr	Rotation-Invariant and Relation-Aware Cross-Domain Adaptation Object Detection Network for Optical Remote Sensing Images
title_full_unstemmed	Rotation-Invariant and Relation-Aware Cross-Domain Adaptation Object Detection Network for Optical Remote Sensing Images
title_sort	rotation-invariant and relation-aware cross-domain adaptation object detection network for optical remote sensing images
publisher	MDPI AG
publishDate	2021
url	https://doaj.org/article/ed656ef591a44896bcc2c5286e5cf6fc
work_keys_str_mv	AT yingchen rotationinvariantandrelationawarecrossdomainadaptationobjectdetectionnetworkforopticalremotesensingimages AT qiliu rotationinvariantandrelationawarecrossdomainadaptationobjectdetectionnetworkforopticalremotesensingimages AT tengwang rotationinvariantandrelationawarecrossdomainadaptationobjectdetectionnetworkforopticalremotesensingimages AT binwang rotationinvariantandrelationawarecrossdomainadaptationobjectdetectionnetworkforopticalremotesensingimages AT xiaoliangmeng rotationinvariantandrelationawarecrossdomainadaptationobjectdetectionnetworkforopticalremotesensingimages
_version_	1718431679316492288

Rotation-Invariant and Relation-Aware Cross-Domain Adaptation Object Detection Network for Optical Remote Sensing Images

Ejemplares similares