Cited time in webofscience Cited time in scopus

Full metadata record

DC Field Value Language
dc.contributor.advisor 최지환 - Younghwan Na - 2022-07-07T02:28:59Z - 2022-07-07T02:28:59Z - 2021 -
dc.identifier.uri en_US
dc.identifier.uri -
dc.description.abstract Semantic segmentation models based on convolutional neural networks (CNNs) have gained much attention in relation to remote sensing and have achieved remarkable performance for the extraction of buildings from high-resolution aerial images. However, the issue of limited generalization for unseen images remains. When there is a domain gap between the training and test datasets, CNN-based segmentation models trained by a training dataset fail to segment buildings for the test dataset. In this paper, we propose segmentation networks based on a domain adaptive transfer attack (DATA) scheme for building extraction from aerial images. The proposed system combines the domain transfer and adversarial attack concepts. Based on the DATA scheme, the distribution of the input images can be shifted to that of the target images while turning images into adversarial examples against a target network. Defending adversarial examples adapted to the target domain can overcome the performance degradation due to the domain gap and increase the robustness of the segmentation model. Cross-dataset experiments and the ablation study are conducted for the three different datasets: the Inria aerial image labeling dataset, the Massachusetts building dataset, and the WHU East Asia dataset. Compared to the performance of the segmentation network without the DATA scheme, the proposed method shows improvements in the overall IoU. Moreover, it is verified that the proposed method outperforms even when compared to feature adaptation (FA) and output space adaptation (OSA). -
dc.description.statementofresponsibility Y -
dc.description.tableofcontents Ⅰ. Introduction 1
1.1. Semantic segmentation of aerial images 1
1.2. Challenging issue in CNN-based segmentation models 2
1.3. Overview of proposed scheme 2
Ⅱ. Basic architecture for semantic segmentation 4
2.1. Semantic segmentation systems 4
2.2. Inria aerial image labeling dataset 6
2.3. Training setup 6
2.4. Test results and comparison with other architectures 7
2.5. Test results with other datasets 8
2.6. Related work for domain adaptation 11
Ⅲ. Domain adaptive transfer attack (DATA) 11
3.1. Overview of the proposed model 12
3.2. Objective function for the generator 14
3.3. Objective function for discriminator 15
3.4. Training the adversarial attack model & discriminator 16
Ⅳ. DATA-based adversarial training and results 19
4.1. Adversary training setup 19
4.2. Comparison with Other Methods 21
4.3. Extended experiments in various environments 24
Ⅴ. Conclusion 26
References 28
dc.format.extent 35 -
dc.language eng -
dc.publisher DGIST -
dc.subject Adversarial network, building extraction, semantic segmentation, domain adaptation, 이미지 분할, 도메인 적응, 적대적 신경망, 건물 추출 -
dc.title Domain Adaptation for Semantic Segmentation of Aerial Images with Adversarial Attack -
dc.title.alternative 적대적 공격을 이용한 항공 이미지 분할에 대한 도메인 적응 -
dc.type Thesis -
dc.identifier.doi 10.22677/thesis.200000361817 -
dc.description.alternativeAbstract 센서 기술의 발달로 다양한 원격 감지 이미지 사용이 가능해졌다. 항공 이미지를 활용하기 위해서는 관심 대상을 추출해서 분할해야 한다. 최근에는 컨볼루션 신경망 (CNN)을 기반으로 한 알고리즘이 활발히 연구되고 있고 항공 이미지 분할에서 뛰어난 성능을 보여준다. 하지만 제한된 일반화 능력에 대한 문제가 여전히 존재한다. 학습 데이터가 목표 데이터와 유사하지 않다면 딥러닝 모델은 목표 데이터세트에서 객체를 분할하지 못한다. 본 논문은 제한된 일반화 문제를 해결하기 위해 적대적 공격 기반 도메인 적응 방법 (Domain Adaptive Transfer Attack, DATA)을 제안한다. 제안된 시스템은 도메인 이전과 적대적 공격 개념을 결합한다. DATA 계획법은 입력 이미지의 분포를 목적 이미지의 분포로 이동시키면서 이미지를 분할 네트워크에 대한 적대적인 예제로 전환할 수 있다. 목적 도메인 쪽으로 옮겨진 적대적인 예제를 방어함으로써 분할 네트워크는 도메인 차이로 인한 성능 저하를 극복한다. 제안한 기법은 세 가지 데이터세트 (Inria aerial image la-beling dataset, Massachusetts building dataset, WHU East Asia dataset)에 대한 교차 실험에서 그 효과를 입증했다. - Master -
dc.contributor.department Information and Communication Engineering -
dc.contributor.coadvisor Sunghoon Im - 2021/02 -
dc.publisher.location Daegu -
dc.description.database dCollection -
dc.citation XT.IM 나64 202102 -
dc.contributor.alternativeDepartment 정보통신융합전공 -
dc.contributor.affiliatedAuthor Younghwan Na -
dc.contributor.affiliatedAuthor Jihwan Choi -
dc.contributor.affiliatedAuthor Sunghoon Im -
dc.contributor.alternativeName 나영환 -
dc.contributor.alternativeName Jihwan Choi -
dc.contributor.alternativeName 임성훈 -
Files in This Item:


기타 데이터 / 1.81 MB / Adobe PDF download
Appears in Collections:
Department of Electrical Engineering and Computer Science Theses Master


  • twitter
  • facebook
  • mendeley

Items in Repository are protected by copyright, with all rights reserved, unless otherwise indicated.