scholarly journals DEEP DOMAIN ADAPTATION BY WEIGHTED ENTROPY MINIMIZATION FOR THE CLASSIFICATION OF AERIAL IMAGES

2020 ◽  
Vol V-2-2020 ◽  
pp. 591-598
Author(s):  
D. Wittich
Keyword(s):  
Data Augmentation ◽  
Domain Adaptation ◽  
Aerial Images ◽  
Target Domain ◽  
Source Domain ◽  
Entropy Minimization ◽  
Training Samples ◽  
Order Of Magnitude ◽  
Weighted Entropy

Abstract. Fully convolutional neural networks (FCN) are successfully used for the automated pixel-wise classification of aerial images and possibly additional data. However, they require many labelled training samples to perform well. One approach addressing this issue is semi-supervised domain adaptation (SSDA). Here, labelled training samples from a source domain and unlabelled samples from a target domain are used jointly to obtain a target domain classifier, without requiring any labelled samples from the target domain. In this paper, a two-step approach for SSDA is proposed. The first step corresponds to a supervised training on the source domain, making use of strong data augmentation to increase the initial performance on the target domain. Secondly, the model is adapted by entropy minimization using a novel weighting strategy. The approach is evaluated on the basis of five domains, corresponding to five cities. Several training variants and adaptation scenarios are tested, indicating that proper data augmentation can already improve the initial target domain performance significantly resulting in an average overall accuracy of 77.5%. The weighted entropy minimization improves the overall accuracy on the target domains in 19 out of 20 scenarios on average by 1.8%. In all experiments a novel FCN architecture is used that yields results comparable to those of the best-performing models on the ISPRS labelling challenge while having an order of magnitude fewer parameters than commonly used FCNs.

scholarly journals ITERATIVE RE-WEIGHTED INSTANCE TRANSFER FOR DOMAIN ADAPTATION

2016 ◽  
Vol III-3 ◽  
pp. 339-346 ◽  
Author(s):  
A. Paul ◽  
F. Rottensteiner ◽  
C. Heipke
Keyword(s):  
Domain Adaptation ◽  
Training Data ◽  
Aerial Images ◽  
Target Domain ◽  
Data Set ◽  
Source Domain ◽  
The Difference ◽  
A New Technique ◽  
Class Labels ◽  
Decision Boundaries

Domain adaptation techniques in transfer learning try to reduce the amount of training data required for classification by adapting a classifier trained on samples from a source domain to a new data set (target domain) where the features may have different distributions. In this paper, we propose a new technique for domain adaptation based on logistic regression. Starting with a classifier trained on training data from the source domain, we iteratively include target domain samples for which class labels have been obtained from the current state of the classifier, while at the same time removing source domain samples. In each iteration the classifier is re-trained, so that the decision boundaries are slowly transferred to the distribution of the target features. To make the transfer procedure more robust we introduce weights as a function of distance from the decision boundary and a new way of regularisation. Our methodology is evaluated using a benchmark data set consisting of aerial images and digital surface models. The experimental results show that in the majority of cases our domain adaptation approach can lead to an improvement of the classification accuracy without additional training data, but also indicate remaining problems if the difference in the feature distributions becomes too large.

scholarly journals ADVERSARIAL DOMAIN ADAPTATION FOR THE CLASSIFICATION OF AERIAL IMAGES AND HEIGHT DATA USING CONVOLUTIONAL NEURAL NETWORKS

2019 ◽  
Vol IV-2/W7 ◽  
pp. 197-204 ◽  
Author(s):  
D. Wittich ◽  
F. Rottensteiner
Keyword(s):  
Neural Networks ◽  
Convolutional Neural Networks ◽  
Domain Adaptation ◽  
Feature Representation ◽  
Training Data ◽  
Aerial Images ◽  
Remotely Sensed Data ◽  
Target Domain ◽  
Deep Convolutional Neural Networks

<p><strong>Abstract.</strong> Domain adaptation (DA) can drastically decrease the amount of training data needed to obtain good classification models by leveraging available data from a source domain for the classification of a new (target) domains. In this paper, we address deep DA, i.e. DA with deep convolutional neural networks (CNN), a problem that has not been addressed frequently in remote sensing. We present a new method for semi-supervised DA for the task of pixel-based classification by a CNN. After proposing an encoder-decoder-based fully convolutional neural network (FCN), we adapt a method for adversarial discriminative DA to be applicable to the pixel-based classification of remotely sensed data based on this network. It tries to learn a feature representation that is domain invariant; domain-invariance is measured by a classifier’s incapability of predicting from which domain a sample was generated. We evaluate our FCN on the ISPRS labelling challenge, showing that it is close to the best-performing models. DA is evaluated on the basis of three domains. We compare different network configurations and perform the representation transfer at different layers of the network. We show that when using a proper layer for adaptation, our method achieves a positive transfer and thus an improved classification accuracy in the target domain for all evaluated combinations of source and target domains.</p>

scholarly journals ITERATIVE RE-WEIGHTED INSTANCE TRANSFER FOR DOMAIN ADAPTATION

2016 ◽  
Vol III-3 ◽  
pp. 339-346 ◽  
Author(s):  
A. Paul ◽  
F. Rottensteiner ◽  
C. Heipke
Keyword(s):  
Domain Adaptation ◽  
Training Data ◽  
Aerial Images ◽  
Target Domain ◽  
Data Set ◽  
Source Domain ◽  
The Difference ◽  
A New Technique ◽  
Class Labels ◽  
Decision Boundaries

Domain adaptation techniques in transfer learning try to reduce the amount of training data required for classification by adapting a classifier trained on samples from a source domain to a new data set (target domain) where the features may have different distributions. In this paper, we propose a new technique for domain adaptation based on logistic regression. Starting with a classifier trained on training data from the source domain, we iteratively include target domain samples for which class labels have been obtained from the current state of the classifier, while at the same time removing source domain samples. In each iteration the classifier is re-trained, so that the decision boundaries are slowly transferred to the distribution of the target features. To make the transfer procedure more robust we introduce weights as a function of distance from the decision boundary and a new way of regularisation. Our methodology is evaluated using a benchmark data set consisting of aerial images and digital surface models. The experimental results show that in the majority of cases our domain adaptation approach can lead to an improvement of the classification accuracy without additional training data, but also indicate remaining problems if the difference in the feature distributions becomes too large.

scholarly journals Category-Sensitive Domain Adaptation for Land Cover Mapping in Aerial Scenes

2019 ◽  
Vol 11 (22) ◽  
pp. 2631 ◽  
Author(s):  
Bo Fang ◽  
Rong Kou ◽  
Li Pan ◽  
Pengfei Chen
Keyword(s):  
Land Cover ◽  
Domain Adaptation ◽  
Feature Space ◽  
Aerial Images ◽  
Land Cover Mapping ◽  
Target Domain ◽  
Generative Adversarial Network ◽  
Source Domain ◽  
Adversarial Network ◽  
Semantic Labeling

Since manually labeling aerial images for pixel-level classification is expensive and time-consuming, developing strategies for land cover mapping without reference labels is essential and meaningful. As an efficient solution for this issue, domain adaptation has been widely utilized in numerous semantic labeling-based applications. However, current approaches generally pursue the marginal distribution alignment between the source and target features and ignore the category-level alignment. Therefore, directly applying them to land cover mapping leads to unsatisfactory performance in the target domain. In our research, to address this problem, we embed a geometry-consistent generative adversarial network (GcGAN) into a co-training adversarial learning network (CtALN), and then develop a category-sensitive domain adaptation (CsDA) method for land cover mapping using very-high-resolution (VHR) optical aerial images. The GcGAN aims to eliminate the domain discrepancies between labeled and unlabeled images while retaining their intrinsic land cover information by translating the features of the labeled images from the source domain to the target domain. Meanwhile, the CtALN aims to learn a semantic labeling model in the target domain with the translated features and corresponding reference labels. By training this hybrid framework, our method learns to distill knowledge from the source domain and transfers it to the target domain, while preserving not only global domain consistency, but also category-level consistency between labeled and unlabeled images in the feature space. The experimental results between two airborne benchmark datasets and the comparison with other state-of-the-art methods verify the robustness and superiority of our proposed CsDA.

scholarly journals A COMPARISON OF TWO STRATEGIES FOR AVOIDING NEGATIVE TRANSFER IN DOMAIN ADAPTATION BASED ON LOGISTIC REGRESSION

2018 ◽  
Vol XLII-2 ◽  
pp. 845-852 ◽  
Author(s):  
A. Paul ◽  
K. Vogt ◽  
F. Rottensteiner ◽  
J. Ostermann ◽  
C. Heipke
Keyword(s):  
Negative Transfer ◽  
Domain Adaptation ◽  
Classification Performance ◽  
Target Domain ◽  
Maximum Mean Discrepancy ◽  
Source Domain ◽  
Training Samples ◽  
Benchmark Datasets ◽  
Consistent Performance ◽  
Class Labels

In this paper we deal with the problem of measuring the similarity between training and tests datasets in the context of transfer learning (TL) for image classification. TL tries to transfer knowledge from a source domain, where labelled training samples are abundant but the data may follow a different distribution, to a target domain, where labelled training samples are scarce or even unavailable, assuming that the domains are related. Thus, the requirements w.r.t. the availability of labelled training samples in the target domain are reduced. In particular, if no labelled target data are available, it is inherently difficult to find a robust measure of relatedness between the source and target domains. This is of crucial importance for the performance of TL, because the knowledge transfer between unrelated data may lead to negative transfer, i.e. to a decrease of classification performance after transfer. We address the problem of measuring the relatedness between source and target datasets and investigate three different strategies to predict and, consequently, to avoid negative transfer in this paper. The first strategy is based on circular validation. The second strategy relies on the Maximum Mean Discrepancy (MMD) similarity metric, whereas the third one is an extension of MMD which incorporates the knowledge about the class labels in the source domain. Our method is evaluated using two different benchmark datasets. The experiments highlight the strengths and weaknesses of the investigated methods. We also show that it is possible to reduce the amount of negative transfer using these strategies for a TL method and to generate a consistent performance improvement over the whole dataset.

scholarly journals NEIGHBOUR-BASED DOMAIN ADAPTATION FOR INVESTIGATION OF TRANSFERABLE ABILITY OF PREVIOUSLY LABELED DATA FOR LAND-COVER CLASSIFICATION OF AERIAL IMAGES

2020 ◽  
Vol XLIII-B2-2020 ◽  
pp. 1329-1335
Author(s):  
R. Zhu ◽  
L. Yan
Keyword(s):  
Land Cover ◽  
Domain Adaptation ◽  
Land Cover Classification ◽  
Classification Performance ◽  
Aerial Images ◽  
Learning Performance ◽  
Training Dataset ◽  
Training Samples ◽  
Target Center

Abstract. Existing land-cover classification methods are usually based on adequate labelled data. But annotating enough training samples is hard and time-consuming. Therefore, we need to investigate how existing labelled data can help to increase land-cover classification. Source labelled data are proposed to be selected by calculating the target center of reliable target pseudo-labelled data for each class in this paper. Then we augment the training dataset with reliable target pesudo-labeled data and selected source labelled data to improve the quality and quantity of training dataset. We also investigate the amount of source labelled data that should be selected and the number of limited target labelled data that can produce good transfer learning performance. The UC Merced dataset is employed as the target dataset to evaluate the proposed approach while the NWPU-RESISC45 dataset is considered as the source labelled data. The experimental results show that selected source labelled data and reliable target pesudo-labeled data may improve the land-cover classification performance if selected source labelled data and reliable target pesudo-labeled data are augmented with the limited target labelled data respectively.

Correlation alignment with attention mechanism for unsupervised domain adaptation

2021 ◽  
pp. 1-7
Author(s):  
Rong Chen ◽  
Chongguang Ren
Keyword(s):  
Natural Language Processing ◽  
Language Processing ◽  
Transfer Process ◽  
Negative Transfer ◽  
Domain Adaptation ◽  
Attention Mechanism ◽  
Target Domain ◽  
Source Domain ◽  
Second Order Statistics ◽  
Unsupervised Domain Adaptation

Domain adaptation aims to solve the problems of lacking labels. Most existing works of domain adaptation mainly focus on aligning the feature distributions between the source and target domain. However, in the field of Natural Language Processing, some of the words in different domains convey different sentiment. Thus not all features of the source domain should be transferred, and it would cause negative transfer when aligning the untransferable features. To address this issue, we propose a Correlation Alignment with Attention mechanism for unsupervised Domain Adaptation (CAADA) model. In the model, an attention mechanism is introduced into the transfer process for domain adaptation, which can capture the positively transferable features in source and target domain. Moreover, the CORrelation ALignment (CORAL) loss is utilized to minimize the domain discrepancy by aligning the second-order statistics of the positively transferable features extracted by the attention mechanism. Extensive experiments on the Amazon review dataset demonstrate the effectiveness of CAADA method.

DAFI

2021 ◽  
Vol 5 (4) ◽  
pp. 1-21
Author(s):  
Hang Li ◽  
Xi Chen ◽  
Ju Wang ◽  
Di Wu ◽  
Xue Liu
Keyword(s):  
Indoor Localization ◽  
Domain Adaptation ◽  
Target Domain ◽  
Localization Accuracy ◽  
Source Domain ◽  
Fine Grained ◽  
Time Signals ◽  
Environmental Variations ◽  
Device Free ◽  
Localization Approach

WiFi-based Device-free Passive (DfP) indoor localization systems liberate their users from carrying dedicated sensors or smartphones, and thus provide a non-intrusive and pleasant experience. Although existing fingerprint-based systems achieve sub-meter-level localization accuracy by training location classifiers/regressors on WiFi signal fingerprints, they are usually vulnerable to small variations in an environment. A daily change, e.g., displacement of a chair, may cause a big inconsistency between the recorded fingerprints and the real-time signals, leading to significant localization errors. In this paper, we introduce a Domain Adaptation WiFi (DAFI) localization approach to address the problem. DAFI formulates this fingerprint inconsistency issue as a domain adaptation problem, where the original environment is the source domain and the changed environment is the target domain. Directly applying existing domain adaptation methods to our specific problem is challenging, since it is generally hard to distinguish the variations in the different WiFi domains (i.e., signal changes caused by different environmental variations). DAFI embraces the following techniques to tackle this challenge. 1) DAFI aligns both marginal and conditional distributions of features in different domains. 2) Inside the target domain, DAFI squeezes the marginal distribution of every class to be more concentrated at its center. 3) Between two domains, DAFI conducts fine-grained alignment by forcing every target-domain class to better align with its source-domain counterpart. By doing these, DAFI outperforms the state of the art by up to 14.2% in real-world experiments.

scholarly journals Joint Partial Optimal Transport for Open Set Domain Adaptation

2020 ◽  
Author(s):  
Renjun Xu ◽  
Pelen Liu ◽  
Yin Zhang ◽  
Fang Cai ◽  
Jindong Wang ◽  
...  
Keyword(s):  
Optimal Transport ◽  
Transport Model ◽  
Domain Adaptation ◽  
General Setting ◽  
Feature Space ◽  
Set Domain ◽  
Target Domain ◽  
Source Domain ◽  
Open Set ◽  
Proposed Model

Domain adaptation (DA) has achieved a resounding success to learn a good classifier by leveraging labeled data from a source domain to adapt to an unlabeled target domain. However, in a general setting when the target domain contains classes that are never observed in the source domain, namely in Open Set Domain Adaptation (OSDA), existing DA methods failed to work because of the interference of the extra unknown classes. This is a much more challenging problem, since it can easily result in negative transfer due to the mismatch between the unknown and known classes. Existing researches are susceptible to misclassification when target domain unknown samples in the feature space distributed near the decision boundary learned from the labeled source domain. To overcome this, we propose Joint Partial Optimal Transport (JPOT), fully utilizing information of not only the labeled source domain but also the discriminative representation of unknown class in the target domain. The proposed joint discriminative prototypical compactness loss can not only achieve intra-class compactness and inter-class separability, but also estimate the mean and variance of the unknown class through backpropagation, which remains intractable for previous methods due to the blindness about the structure of the unknown classes. To our best knowledge, this is the first optimal transport model for OSDA. Extensive experiments demonstrate that our proposed model can significantly boost the performance of open set domain adaptation on standard DA datasets.

scholarly journals Unsupervised Domain Adaptation by Statistics Alignment for Deep Sleep Staging Networks

2021 ◽  
Author(s):  
Jiahao Fan ◽  
Hangyu Zhu ◽  
Xinyu Jiang ◽  
Long Meng ◽  
Cong Fu ◽  
...  
Keyword(s):  
Large Scale ◽  
Domain Adaptation ◽  
Source Model ◽  
Deep Sleep ◽  
Sleep Staging ◽  
Target Domain ◽  
Source Domain ◽  
Unsupervised Domain Adaptation ◽  
Generalization Problem ◽  
Source Models

Deep sleep staging networks have reached top performance on large-scale datasets. However, these models perform poorer when training and testing on small sleep cohorts due to data inefficiency. Transferring well-trained models from large-scale datasets (source domain) to small sleep cohorts (target domain) is a promising solution but still remains challenging due to the domain-shift issue. In this work, an unsupervised domain adaptation approach, domain statistics alignment (DSA), is developed to bridge the gap between the data distribution of source and target domains. DSA adapts the source models on the target domain by modulating the domain-specific statistics of deep features stored in the Batch Normalization (BN) layers. Furthermore, we have extended DSA by introducing cross-domain statistics in each BN layer to perform DSA adaptively (AdaDSA). The proposed methods merely need the well-trained source model without access to the source data, which may be proprietary and inaccessible. DSA and AdaDSA are universally applicable to various deep sleep staging networks that have BN layers. We have validated the proposed methods by extensive experiments on two state-of-the-art deep sleep staging networks, DeepSleepNet+ and U-time. The performance was evaluated by conducting various transfer tasks on six sleep databases, including two large-scale databases, MASS and SHHS, as the source domain, four small sleep databases as the target domain. Thereinto, clinical sleep records acquired in Huashan Hospital, Shanghai, were used. The results show that both DSA and AdaDSA could significantly improve the performance of source models on target domains, providing novel insights into the domain generalization problem in sleep staging tasks.<br>

Sign in / Sign up

Export Citation Format

Share Document