Unlabeled Target Domain Research Articles

Reducing Domain Gap in Frequency and Spatial Domain for Cross-Modality Domain Adaptation on Medical Image Segmentation

Unsupervised domain adaptation (UDA) aims to learn a model trained on source domain and performs well on unlabeled target domain. In medical image segmentation field, most existing UDA methods depend on adversarial learning to address the domain gap between different image modalities, which is ineffective due to its complicated training process. In this paper, we propose a simple yet effective UDA method based on frequency and spatial domain transfer under multi-teacher distillation framework. In the frequency domain, we first introduce non-subsampled contourlet transform for identifying domain-invariant and domain-variant frequency components (DIFs and DVFs), and then keep the DIFs unchanged while replacing the DVFs of the source domain images with that of the target domain images to narrow the domain gap. In the spatial domain, we propose a batch momentum update-based histogram matching strategy to reduce the domain-variant image style bias. Experiments on two commonly used cross-modality medical image segmentation datasets show that our proposed method achieves superior performance compared to state-of-the-art methods.

Decorate the Newcomers: Visual Domain Prompt for Continual Test Time Adaptation

Continual Test-Time Adaptation (CTTA) aims to adapt the source model to continually changing unlabeled target domains without access to the source data. Existing methods mainly focus on model-based adaptation in a self-training manner, such as predicting pseudo labels for new domain datasets. Since pseudo labels are noisy and unreliable, these methods suffer from catastrophic forgetting and error accumulation when dealing with dynamic data distributions. Motivated by the prompt learning in NLP, in this paper, we propose to learn an image-layer visual domain prompt for target domains while having the source model parameters frozen. During testing, the changing target datasets can be adapted to the source model by reformulating the input data with the learned visual prompts. Specifically, we devise two types of prompts, i.e., domains-specific prompts and domains-agnostic prompts, to extract current domain knowledge and maintain the domain-shared knowledge in the continual adaptation. Furthermore, we design a homeostasis-based adaptation strategy to suppress domain-sensitive parameters in domain-invariant prompts to learn domain-shared knowledge more effectively. This transition from the model-dependent paradigm to the model-free one enables us to bypass the catastrophic forgetting and error accumulation problems. Experiments show that our proposed method achieves significant performance gains over state-of-the-art methods on four widely-used benchmarks, including CIFAR-10C, CIFAR-100C, ImageNet-C, and VLCS datasets.

Prototypical Partial Optimal Transport for Universal Domain Adaptation

Universal domain adaptation (UniDA) aims to transfer knowledge from a labeled source domain to an unlabeled target domain without requiring the same label sets of both domains. The existence of domain and category shift makes the task challenging and requires us to distinguish “known” samples (i.e., samples whose labels exist in both domains) and “unknown” samples (i.e., samples whose labels exist in only one domain) in both domains before reducing the domain gap. In this paper, we consider the problem from the point of view of distribution matching which we only need to align two distributions partially. A novel approach, dubbed mini-batch Prototypical Partial Optimal Transport (m-PPOT), is proposed to conduct partial distribution alignment for UniDA. In training phase, besides minimizing m-PPOT, we also leverage the transport plan of m-PPOT to reweight source prototypes and target samples, and design reweighted entropy loss and reweighted cross-entropy loss to distinguish “known” and “unknown” samples. Experiments on four benchmarks show that our method outperforms the previous state-of-the-art UniDA methods.

Discriminability and Transferability Estimation: A Bayesian Source Importance Estimation Approach for Multi-Source-Free Domain Adaptation

Source free domain adaptation (SFDA) transfers a single-source model to the unlabeled target domain without accessing the source data. With the intelligence development of various fields, a zoo of source models is more commonly available, arising in a new setting called multi-source-free domain adaptation (MSFDA). We find that the critical inborn challenge of MSFDA is how to estimate the importance (contribution) of each source model. In this paper, we shed new Bayesian light on the fact that the posterior probability of source importance connects to discriminability and transferability. We propose Discriminability And Transferability Estimation (DATE), a universal solution for source importance estimation. Specifically, a proxy discriminability perception module equips with habitat uncertainty and density to evaluate each sample's surrounding environment. A source-similarity transferability perception module quantifies the data distribution similarity and encourages the transferability to be reasonably distributed with a domain diversity loss. Extensive experiments show that DATE can precisely and objectively estimate the source importance and outperform prior arts by non-trivial margins. Moreover, experiments demonstrate that DATE can take the most popular SFDA networks as backbones and make them become advanced MSFDA solutions.

Adversarial Alignment for Source Free Object Detection

Source-free object detection (SFOD) aims to transfer a detector pre-trained on a label-rich source domain to an unlabeled target domain without seeing source data. While most existing SFOD methods generate pseudo labels via a source-pretrained model to guide training, these pseudo labels usually contain high noises due to heavy domain discrepancy. In order to obtain better pseudo supervisions, we divide the target domain into source-similar and source-dissimilar parts and align them in the feature space by adversarial learning.Specifically, we design a detection variance-based criterion to divide the target domain. This criterion is motivated by a finding that larger detection variances denote higher recall and larger similarity to the source domain. Then we incorporate an adversarial module into a mean teacher framework to drive the feature spaces of these two subsets indistinguishable. Extensive experiments on multiple cross-domain object detection datasets demonstrate that our proposed method consistently outperforms the compared SFOD methods. Our implementation is available at https://github.com/ChuQiaosong.

Distributionally robust unsupervised domain adaptation

Obtaining ground-truth label information from real-world data along with uncertainty quantification can be difficult or even infeasible. In the absence of labeled data for a certain task, unsupervised domain adaptation (UDA) techniques have shown great accomplishment by learning transferable knowledge from labeled source domain data and applying it to unlabeled target domain data, yet scarce studies consider addressing uncertainties under domain shifts to improve the model robustness. Distributionally robust learning (DRL) is emerging as a high-potential technique for building reliable learning systems that are robust to distribution shifts. In this study, we propose a distributionally robust unsupervised domain adaptation (DRUDA) method to enhance the machine learning model generalization ability under input space perturbations. The DRL-based UDA learning scheme is formulated as a min–max optimization problem by optimizing worst-case perturbations of the training source data. Our Wasserstein distributionally robust framework can reduce the shifts in the joint distributions across domains. The proposed DRUDA has been tested on digit datasets and the Office-31 dataset, compared with other state-of-the-art domain adaptation techniques. Our experimental results show that the proposed DRUDA leads to improvements in domain adaptation accuracy performance on target domains.

Unsupervised Cross-modality Adaptation via Dual Structural-Oriented Guidance for 3D Medical Image Segmentation.

Deep convolutional neural networks (CNNs) have achieved impressive performance in medical image segmentation; however, their performance could degrade significantly when being deployed to unseen data with heterogeneous characteristics. Unsupervised domain adaptation (UDA) is a promising solution to tackle this problem. In this work, we present a novel UDA method, named dual adaptation-guiding network (DAG-Net), which incorporates two highly effective and complementary structural-oriented guidance in training to collaboratively adapt a segmentation model from a labelled source domain to an unlabeled target domain. Specifically, our DAG-Net consists of two core modules: 1) Fourier-based contrastive style augmentation (FCSA) which implicitly guides the segmentation network to focus on learning modality-insensitive and structural-relevant features, and 2) residual space alignment (RSA) which provides explicit guidance to enhance the geometric continuity of the prediction in the target modality based on a 3D prior of inter-slice correlation. We have extensively evaluated our method with cardiac substructure and abdominal multi-organ segmentation for bidirectional cross-modality adaptation between MRI and CT images. Experimental results on two different tasks demonstrate that our DAG-Net greatly outperforms the state-of-the-art UDA approaches for 3D medical image segmentation on unlabeled target images.

A Novel Cluster Matching-Based Improved Kernel Fisher Criterion for Image Classification in Unsupervised Domain Adaptation

Unsupervised domain adaptation (UDA) is a popular approach to reducing distributional discrepancies between labeled source and the unlabeled target domain (TD) in machine learning. However, current UDA approaches often align feature distributions between two domains explicitly without considering the target distribution and intra-domain category information, potentially leading to reduced classifier efficiency when the distribution between training and test sets differs. To address this limitation, we propose a novel approach called Cluster Matching-based Improved Kernel Fisher criterion (CM-IKFC) for object classification in image analysis using machine learning techniques. CM-IKFC generates accurate pseudo-labels for each target sample by considering both domain distributions. Our approach employs K-means clustering to cluster samples in the latent subspace in both domains and then conducts cluster matching in the TD. During the model component training stage, the Improved Kernel Fisher Criterion (IKFC) is presented to extend cluster matching and preserve the semantic structure and class transitions. To further enhance the performance of the Kernel Fisher criterion, we use a normalized parameter, due to the difficulty in solving the characteristic equation that draws inspiration from symmetry theory. The proposed CM-IKFC method minimizes intra-class variability while boosting inter-class variants in all domains. We evaluated our approach on benchmark datasets for UDA tasks and our experimental findings show that CM-IKFC is superior to current state-of-the-art methods.

Addressing multi-site functional MRI heterogeneity through dual-expert collaborative learning for brain disease identification.

Several studies employ multi-site rs-fMRI data for major depressive disorder (MDD) identification, with a specific site as the to-be-analyzed target domain and other site(s) as the source domain. But they usually suffer from significant inter-site heterogeneity caused by the use of different scanners and/or scanning protocols and fail to build generalizable models that can well adapt to multiple target domains. In this article, we propose a dual-expert fMRI harmonization (DFH) framework for automated MDD diagnosis. Our DFH is designed to simultaneously exploit data from a single labeled source domain/site and two unlabeled target domains for mitigating data distribution differences across domains. Specifically, the DFH consists of a domain-generic student model and two domain-specific teacher/expert models that are jointly trained to perform knowledge distillation through a deep collaborative learning module. A student model with strong generalizability is finally derived, which can be well adapted to unseen target domains and analysis of other brain diseases. To the best of our knowledge, this is among the first attempts to investigate multi-target fMRI harmonization for MDD diagnosis. Comprehensive experiments on 836 subjects with rs-fMRI data from 3 different sites show the superiority of our method. The discriminative brain functional connectivities identified by our method could be regarded as potential biomarkers for fMRI-related MDD diagnosis.

Transferable adversarial masked self-distillation for unsupervised domain adaptation

Unsupervised domain adaptation (UDA) aims to transfer knowledge from a labeled source domain to a related unlabeled target domain. Most existing works focus on minimizing the domain discrepancy to learn global domain-invariant representation using CNN-based architecture while ignoring both transferable and discriminative local representation, e.g, pixel-level and patch-level representation. In this paper, we propose the Transferable Adversarial Masked Self-distillation based on Vision Transformer architecture to enhance the transferability of UDA, named TAMS. Specifically, TAMS jointly optimizes three objectives to learn both task-specific class-level global representation and domain-specific local representation. First, we introduce adversarial masked self-distillation objective to distill representation from a full image to the representation predicted from a masked image, which aims to learn task-specific global class-level representation. Second, we introduce masked image modeling objectives to learn local pixel-level representation. Third, we introduce an adversarial weighted cross-domain adaptation objective to capture discriminative potentials of patch tokens, which aims to learn both transferable and discriminative domain-specific patch-level representation. Extensive studies on four benchmarks and the experimental results show that our proposed method can achieve remarkable improvements compared to previous state-of-the-art UDA methods.

Class-specific and self-learning local manifold structure for domain adaptation

Domain adaptation (DA) is a powerful technology that allows a classifier trained on a well-labeled source domain to be adapted to an unlabeled target domain with different distributions. Existing DA methods aim to enhance feature transferability by reducing distribution distance, and they often rely on preserving the global discriminative (GD) structure to boost feature discriminability. However, strict GD loss may degrade transferability, and poor discriminability may result from wrongly labeled samples. To address these issues, we propose a new approach called class-specific and self-learning local manifold structure (CSSL-LM), to extract more desirable features for better DA effects. Specifically, it draws two data points close with large weight if they are from the same class and close in the original feature space. This approach is more relaxed than GD and thus mitigates the negative effect on transferability. Moreover, CSSL-LM is more robust to wrongly labeled samples than GD since two data points that are wrongly labeled as the same classes have small weight and are not required to be close. Inspired by previous adaptive local manifold learning, we utilize a self-learning mechanism to model CSSL-LM more accurately and reliably, particularly for feature-corrupted samples. Extensive experiments on four DA benchmarks verify that CSSL-LM outperforms some state-of-the-art methods. We also construct DA datasets that are randomly wrongly labeled or feature-corrupted to further evaluate CSSL-LM’s robustness.

Multi-Source Collaborative Contrastive Learning for Decentralized Domain Adaptation

Unsupervised multi-source domain adaptation aims to obtain a model working well on the unlabeled target domain by reducing the domain gap between the labeled source domains and the unlabeled target domain. Considering the data privacy and storage cost, data from multiple source domains and target domain are isolated and decentralized. This data decentralization scenario brings the difficulty of domain alignment for reducing the domain gap between the decentralized source domains and target domain, respectively. For conducting domain alignment under the data decentralization scenario, we propose Multi-source Collaborative Contrastive learning for decentralized Domain Adaptation (MCC-DA). The models from other domains are used as the bridge to reduce the domain gap. On the source domains and target domain, we penalize the inconsistency of data features extracted from the source domain models and target domain model by contrastive alignment. With the collaboration of source domain models and target domain model, the domain gap between decentralized source domains and target domain is reduced without accessing the data from other domains. The experiment results on multiple benchmarks indicate that our method can reduce the domain gap effectively and outperform the state-of-the-art methods significantly.

Deep Joint Distribution Alignment: A Novel Enhanced-Domain Adaptation Mechanism for Fault Transfer Diagnosis.

Various domain adaptation (DA) methods have been proposed to address distribution discrepancy and knowledge transfer between the source and target domains. However, many DA models focus on matching the marginal distributions of two domains and cannot satisfy fault-diagnosed-task requirements. To enhance the ability of DA, a new DA mechanism, called deep joint distribution alignment (DJDA), is proposed to simultaneously reduce the discrepancy in marginal and conditional distributions between two domains. A new statistical metric that can align the means and covariances of two domains is designed to match the marginal distributions of the source and target domains. To align the class conditional distributions, a Gaussian mixture model is used to obtain the distribution of each category in the target domain. Then, the conditional distributions of the source domain are computed via maximum-likelihood estimation, and information entropy and Wasserstein distance are employed to reduce class conditional distribution discrepancy between the two domains. With joint distribution alignment, DJDA can achieve domain confusion to the highest degree. DJDA is applied to the fault transfer diagnosis of a wind turbine gearbox and cross-bearing with unlabeled target-domain samples. Experimental results verify that DJDA outperforms other typical DA models.

Unsupervised Domain Adaptation for Image Classification and Object Detection Using Guided Transfer Learning Approach and JS Divergence.

Unsupervised domain adaptation (UDA) is a transfer learning technique utilized in deep learning. UDA aims to reduce the distribution gap between labeled source and unlabeled target domains by adapting a model through fine-tuning. Typically, UDA approaches assume the same categories in both domains. The effectiveness of transfer learning depends on the degree of similarity between the domains, which determines an efficient fine-tuning strategy. Furthermore, domain-specific tasks generally perform well when the feature distributions of the domains are similar. However, utilizing a trained source model directly in the target domain may not generalize effectively due to domain shift. Domain shift can be caused by intra-class variations, camera sensor variations, background variations, and geographical changes. To address these issues, we design an efficient unsupervised domain adaptation network for image classification and object detection that can learn transferable feature representations and reduce the domain shift problem in a unified network. We propose the guided transfer learning approach to select the layers for fine-tuning the model, which enhances feature transferability and utilizes the JS-Divergence to minimize the domain discrepancy between the domains. We evaluate our proposed approaches using multiple benchmark datasets. Our domain adaptive image classification approach achieves 93.2% accuracy on the Office-31 dataset and 75.3% accuracy on the Office-Home dataset. In addition, our domain adaptive object detection approach achieves 51.1% mAP on the Foggy Cityscapes dataset and 72.7% mAP on the Indian Vehicle dataset. We conduct extensive experiments and ablation studies to demonstrate the effectiveness and efficiency of our work. Experimental results also show that our work significantly outperforms the existing methods.

DeepAstroUDA: semi-supervised universal domain adaptation for cross-survey galaxy morphology classification and anomaly detection

Artificial intelligence methods show great promise in increasing the quality and speed of work with large astronomical datasets, but the high complexity of these methods leads to the extraction of dataset-specific, non-robust features. Therefore, such methods do not generalize well across multiple datasets. We present a universal domain adaptation method, DeepAstroUDA, as an approach to overcome this challenge. This algorithm performs semi-supervised domain adaptation (DA) and can be applied to datasets with different data distributions and class overlaps. Non-overlapping classes can be present in any of the two datasets (the labeled source domain, or the unlabeled target domain), and the method can even be used in the presence of unknown classes. We apply our method to three examples of galaxy morphology classification tasks of different complexities (three-class and ten-class problems), with anomaly detection: (1) datasets created after different numbers of observing years from a single survey (Legacy Survey of Space and Time mock data of one and ten years of observations); (2) data from different surveys (Sloan Digital Sky Survey (SDSS) and DECaLS); and (3) data from observing fields with different depths within one survey (wide field and Stripe 82 deep field of SDSS). For the first time, we demonstrate the successful use of DA between very discrepant observational datasets. DeepAstroUDA is capable of bridging the gap between two astronomical surveys, increasing classification accuracy in both domains (up to on the unlabeled data), and making model performance consistent across datasets. Furthermore, our method also performs well as an anomaly detection algorithm and successfully clusters unknown class samples even in the unlabeled target dataset.

Semi-supervised graph convolutional networks for the domain adaptive recognition of thyroid nodules in cross-device ultrasound images.

Ultrasound plays a critical role in the early screening and diagnosis of cancers. Although deep neural networks have been widely investigated in the computer-aided diagnosis (CAD) of different medical images, diverse ultrasound devices, and image modalities pose challenges for clinical applications, especially in the recognition of thyroid nodules having various shapes and sizes. More generalized and extensible methods need to be developed for the cross-devices recognition of thyroidnodules. In this work, a semi-supervised graph convolutional deep learning framework is proposed for the domain adaptative recognition of thyroid nodules across several ultrasound devices. A deep classification network, trained on a source domain with a specific device, can be transferred to recognize thyroid nodules on the target domain with other devices, using only few manual annotated ultrasoundimages. This study presents a semi-supervised graph-convolutional-network-based domain adaptation framework, namely Semi-GCNs-DA. Based on the ResNet backbone, it is extended in three aspects for domain adaptation, that is, graph convolutional networks (GCNs) for the connection construction between source and target domains, semi-supervised GCNs for accurate target domain recognition, and pseudo labels for unlabeled target domains. Data were collected from 1498 patients comprising 12 108 images with or without thyroid nodules under three different ultrasound devices. Accuracy, Sensitivity and Specificity were used for the performanceevaluation. The proposed method was validated on six groups of data for a single source domain adaptation task, the mean Accuracy was 0.9719± 0.0023, 0.9928± 0.0022, 0.9353± 0.0105, 0.8727± 0.0021, 0.7596± 0.0045, 0.8482± 0.0092, which achieved better performance in comparison with the state-of-the-art. The proposed method was also validated on three groups of multiple source domain adaptation tasks. In particular, when using X60 and HS50 as the source domain data, and H60 as the target domain, it can achieve the Accuracy of 0.8829± 0.0079, Sensitivity of 0.9757± 0.0001, and Specificity of 0.7894± 0.0164. Ablation experiments also demonstrated the effectiveness of the proposedmodules. The developed Semi-GCNs-DA framework can effectively recognize thyroid nodules on different ultrasound devices. The developed semi-supervised GCNs can be further extended to the domain adaptation problems for other modalities of medicalimages.

Multi-Level Progressive Learning for Unsupervised Vehicle Re-Identification

Vehicle re-identification (ReID) technology has played a more and more important role in Intelligent Transport System (ITS), which aims at searching the same query vehicle identity from a large amount of gallery datasets under different non-overlapping camera views. Current related researches mainly focus on discriminative feature mining of vehicle images and train the model in a fully supervised manner which highly relies on the manual annotations of training data. However, it is labor-consuming and impractical to generate the annotation for each sample image in real-word applications especially for those large-scale transport systems with tons of surveillance data. To this point, we propose in this paper a multi-level progressive learning (MLPL) method for unsupervised vehicle ReID, which gives a good performance by only utilizing the unlabeled target domain images. We firstly introduce a multi-branch architecture to explore the vehicle representations in different level, which consists of one branch for global feature and two branches for local feature learning. A density-based clustering method is employed to generate pseudo labels. Combining with the unique model, we propose a novel re-clustering method to better mine the labels with high reliability. Then a dynamic progressive contrast learning (DPCL) strategy is carefully designed to train the network based on these clustered labels. DPCL could dynamically adjust the training process to maximally strengthen the multi-level feature learning. Moreover, we further propose a self-adaptive loss balance method to automatically compute the weights of different losses during each training iteration. Comprehensive experiments are conducted on several mainstream evaluation datasets, including VeRi776, VehicleID and CityFlowV2-ReID. Compared to other existed unsupervised methods, our approach achieves the new state-of-the-art performance.

Unsupervised domain adaptation for regression using dictionary learning

Unsupervised domain adaptation aims to generalize the knowledge learned on a labeled source domain across an unlabeled target domain. Most of existing unsupervised approaches are feature-based methods that seek to find domain invariant features. Despite their wide applications, these approaches proved to have some limitations especially in regression tasks. In this paper, we study the problem of unsupervised domain adaptation for regression tasks. We highlight the obstacles faced in regression compared to a classification task in terms of sensitivity to the scattering of data in feature space. We take this issue and propose a new unsupervised domain adaptation approach based on dictionary learning. We seek to learn a dictionary on source data and follow an optimal direction trajectory to minimize the residue of the reconstruction of the target data with the same dictionary. For stable training of a neural network, we provide a robust implementation of a projected gradient descent dictionary learning framework, which allows to have a backpropagation friendly end-to-end method. Experimental results show that the proposed method outperforms significantly most of state-of-the-art methods on several well-known benchmark datasets, especially when transferring knowledge from synthetic to real domains.

Transformer-Based Domain-Specific Representation for Unsupervised Domain Adaptive Vehicle Re-Identification

Fully-supervised vehicle re-identification (re-ID) methods are faced with performance degradation when applied to new image domains. Therefore, developing unsupervised domain adaptation (UDA) to transfer the knowledge from learned source domain to new unlabeled target domain becomes an indispensable task. It is challenging because different domains have various image appearances, such as different backgrounds, illuminations and resolutions, especially when cameras have different viewpoints. To tackle this domain gap issue, a novel Transformer-based Domain-Specific Representation learning network (TDSR) is proposed to dynamically focus on corresponding detailed hints for each domain. Specifically, with the source and target domain being trained simultaneously, a domain encoding module is proposed to introduce domain information into the network. The original features of source and target domains are enriched with these domain encodings first, and then sequentially processed by a Transformer encoder to model contextual information and a decoder to summarize the encoded features into the final domain-specific feature representations. Moreover, we propose a Contrastive Clustering Loss (CCL) to directly optimize the distribution of features at cluster level. Instances are overall pulled closer to the prototype of the same identity, and pushed farther from the prototypes of different identities. It helps compact the clusters in the latent space and improve the discriminative capability of the network, leading to more accurate pseudo-label assignment in TDSR. Our method outperforms the state-of-the-art UDA methods on vehicle re-ID benchmark datasets VeRi and VehicleID on both real-world to real-world and synthetic to real-world settings.

Source free unsupervised domain adaptation for electro-mechanical actuator fault diagnosis

A common necessity for prior unsupervised domain adaptation methods that can improve the domain adaptation in unlabeled target domain dataset is access to source domain dataset and target domain dataset simultaneously. However, data privacy makes it not always possible to access source domain dataset and target domain dataset in actual industrial equipment simultaneously, especially for aviation component like Electro-Mechanical Actuator (EMA) whose dataset are often not shareable due to the data copyright and confidentiality. To address this problem, this paper proposes a source free unsupervised domain adaptation framework for EMA fault diagnosis. The proposed framework is a combination of feature network and classifier. Firstly, source domain datasets are only applied to train a source model. Secondly, the well-trained source model is transferred to target domain and classifier is frozen based on source domain hypothesis. Thirdly, nearest centroid filtering is introduced to filter the reliable pseudo labels for unlabeled target domain dataset, and finally, supervised learning and pseudo label clustering are applied to fine-tune the transferred model. In comparison with several traditional unsupervised domain adaptation methods, case studies based on low- and high-frequency monitoring signals on EMA indicate the effectiveness of the proposed method.