Feature alignment and spatio-temporal domain adaptive strategy for aeroengine virtual sensor model construction under domain shifts

Building extraction from high-resolution remote sensing imagery holds significant value for urban planning, disaster assessment, and geospatial analysis. However, current semantic segmentation models still face limitations when handling complex scenarios characterized by diverse building morphologies, significant scale variations, and blurred boundaries. To address the challenges of insufficient long-range dependency modeling, suboptimal multi-scale feature representation, and weak spatial adaptability, this paper proposes a building extraction network that integrates multi-scale sequence modeling with spatial adaptive enhancement. Adopting UPerNet (equipped with ConvNeXt-Tiny) as the baseline framework, the proposed method introduces a dedicated PyramidSSM-based neck (PyramidSSMNeck) as the primary design for multi-scale feature alignment and fusion, and further integrates three enhancement components (S6 (SSM-based), LSKNet, and SAFM) that provide additional improvements mainly reflected in boundary delineation. Specifically, PyramidSSMNeck performs structured cross-scale feature projection, alignment, and aggregation to strengthen multi-scale representation; S6 enhances long-range contextual modeling, LSKNet adaptively adjusts spatial receptive fields to accommodate scale variations, and SAFM modulates feature responses with spatial cues to refine boundaries and fine details—forming a unified framework in which PyramidSSMNeck primarily drives multi-scale alignment and fusion, while S6, LSKNet, and SAFM further enhance long-range context modeling and spatial adaptivity, mainly benefiting boundary preservation and fine-detail integrity. Experiments were conducted on the WHU Building, INRIA, and a self-constructed Ganzhou urban dataset, and the results indicate that the proposed method achieved IoU scores of 91.29%, 81.96%, and 88.18% across the three datasets, outperforming the baseline UPerNet (ConvNeXt-Tiny) by 2.37%, 0.88%, and 3.68%, respectively, with F1-scores consistently exceeding 90%. Importantly, ablation results indicate that the majority of region-level gains (IoU/F1) come from PyramidSSMNeck, whereas the additional modules contribute more prominently to boundary quality, yielding a Boundary IoU increase from 63.29% to 65.63% (+2.34) from the neck-only setting to the full model. Visualization results further support the method’s advantages in boundary preservation and detail integrity, and additional cross-domain transfer experiments (zero-shot and few-shot from WHU to Ganzhou) suggest improved robustness under domain shift.

Multitarget domain adaptation (MTDA) is a transfer learning task that uses knowledge extracted from a labeled source domain to adapt across multiple unlabeled target domains. The MTDA setting is more complicated than the single-source-single-target domain adaptation (S3TDA) setting because domain shift not only exists in each pair of a source–target domain but also exists among different target domains. In addition, multiple-target domains have their own unique characteristics because they are often collected from various conditions. The semantic information in each target domain can be damaged when they are naïvely merged into a single-target domain. Therefore, the trained model struggles to distinguish between representations in the combined target domain, which degrades the classification performance. Furthermore, the knowledge transferability from the source domain to multiple-target domains in prior studies leaves room for improvement because they only focus on exploiting the relationship of source–target pairs while failing to consider the correlation among multiple-target domains. This article introduces an easy-to-hard adaption structure to solve these problems in MTDA. The proposed method consists of three components: Extracting source representations, Hierarchical intratarget feature Alignment, and Collaborative intertarget feature Alignment, called EHACA. These components are used to encode the semantic information in each target domain and explore the relationships between the source and target domains, and among different target domains. The proposed method shows outstanding classification performance over five remote sensing datasets of MTDA tasks, surpassing state-of-the-art approaches in most experimental scenarios.

Category-level selective dual-adversarial network using significance-augmented unsupervised domain adaptation for surface defect detection

Efficient implementation of Visible-Infrared Person Re-identification (VI-ReID) is important for intelligent transportation and surveillance systems, where the primary challenge lies in the semantic discrepancy between visible and infrared modalities. Although existing methods have made progress in mitigating modality discrepancies, single-stage feature alignment remains prone to semantic shifts. We argue that hierarchical transitions in the feature alignment process can alleviate this issue. Based on this observation, we propose a Multi-Stage Feature Alignment and Cross-Modality Collaborative Fusion (MS-CF) framework, which performs multi-dimensional modality discrepancy optimization and suppresses redundant information, thereby ensuring cross-modality consistency and intra-modality stability. The MS-CF framework consists of three modules. The Dual-Path Cross-Layer Attention (DCA) module enhances semantic and structural representations of intermediate features. The Balanced Feature Normalization (BFN) module improves feature distribution consistency and discriminability by incorporating modality constraints and feature sparsification. The Multi-Stage Hybrid-Modality Alignment (MS-HMA) strategy applies collaborative constraints to hybrid modalities after initial single-modality alignment, enabling coarse-to-fine semantic convergence. Extensive experiments on the SYSU-MM01 and RegDB datasets demonstrate that the MS-CF framework outperforms state-of-the-art methods. In particular, under the indoor-search setting of SYSU-MM01, MS-CF achieves improvements of 5.75% in Rank-1 accuracy and 3.97% in mAP, validating its effectiveness in enhancing robustness and recognition performance.

This research paper presents the study of modelling and validating the results of simulation of LiDAR sensor with actual sensor data. The model can be further utilized for virtual testing of various ADAS functions. The virtual sensor is a physics-based sensor capable of emulating the real sensor. Initially an in-depth study of Velodyne VLP-16 sensor was carried out in order to parameterize the virtual sensor model. The sensor is modelled virtually in Gazebo simulation environment. Gazebo is used along with ROS (Robotic Operating Interface) in order to record the sensor data. Data conversion from sensor message format to Cartesian coordinates has been carried out. This was done in order to have a common format of actual and virtual data for validation. Four different objects were selected for validating the sensor data. The validation methodology is based on comparison of actual recorded data generated by Velodyne VLP-16 LiDAR and virtually generated data by the virtual LiDAR sensor model. The validation proves that the in case of static scenarios, the virtual lidar sensor model is 99 % accurate in measuring the range to the object. Although there is slight deviation from the actual value, it is in the acceptable limits.

DETR-style detectors stand out amongst in-domain scenarios, but their properties in domain shift settings are under-explored. This paper aims to build a simple but effective baseline with a DETR-style detector on domain shift settings based on two findings. For one, mitigating the domain shift on the backbone and the decoder output features excels in getting favorable results. For another, advanced domain alignment methods in both parts further enhance the performance. Thus, we propose the Object-Aware Alignment (OAA) module and the Optimal Transport based Alignment (OTA) module to achieve comprehensive domain alignment on the outputs of the backbone and the detector. The OAA module aligns the foreground regions identified by pseudo-labels in the backbone outputs, leading to domain-invariant base features. The OTA module utilizes sliced Wasserstein distance to maximize the retention of location information while minimizing the domain gap in the decoder outputs. We implement the findings and the alignment modules into our adaptation method, and it benchmarks the DETR-style detector on the domain shift settings. Experiments on various domain adaptive scenarios validate the effectiveness of our method.

From patch, sample to domain: Capture geometric structures for few-shot learning

Deep learning techniques have achieved significant advancements in medical image segmentation in recent years. However, model generalization remains severely constrained by domain shift, particularly in cross-modal medical image segmentation tasks. Traditional segmentation models struggle to generalize effectively to unseen target domains due to differences in joint probability distributions across medical imaging modalities. Existing methods primarily focus on unsupervised domain adaptation (UDA) and domain generalization (DG) techniques. While UDA methods face practical limitations due to difficulties in obtaining target domain data, current DG approaches often overlook inherent anatomical priors in medical images as well as the heterogeneity and sparsity of lesion regions. To address these challenges, this paper proposes a cross-modal medical image segmentation framework that integrates the Vision Mamba model with dynamic domain generalization. The framework achieves cross-domain feature alignment and multi-scale feature fusion by leveraging bidirectional state-space sequence modeling, Bezier curve-style enhancement, and a dual-normalization strategy. Additionally, the VEBlock module is introduced, which effectively combines the dynamic sequence modeling capabilities of the Mamba model with non-local attention mechanisms to better capture cross-modal global dependencies. Experimental results on the BraTS 2018 and cross-modal cardiac datasets demonstrate significant improvements in cross-modal segmentation. For example, in T2 → T1 tasks, our framework achieves an average Dice score of 56.22%, outperforming baseline methods by 1.78% while reducing Hausdorff distance for tumor boundaries to 13.26mm. Furthermore, in the Cardiac CT → MRI tasks, the Hausdorff distance is optimized to 27.34mm, validating the proposed framework's strong generalization capability for complex anatomical structures.

Traditional deep learning Synthetic Aperture Radar (SAR) image object detection methods fail to provide effective detection results when faced with SAR image datasets with different joint probability distributions obtained from multiple imaging satellites. In this article, an automatic SAR image object detection method based on domain adaptation is proposed to adapt to unlabeled target domain datasets acquired by different satellites. On the basis of introducing an adversarial domain adaptation learning strategy, we propose Adversarial Learning Attention (ALA) and Compensation Loss Module (CLM) on the baseline network. In ALA, considering the great difference in the scattering intensity of SAR images, the entropy vector can be used to distinguish the high-entropy and low-entropy regions among them and assign the corresponding weights, based on which the adversarial domain adaptation learning attention is proposed to achieve instance-level feature alignment and pixel-level feature alignment in source domain and target domain, respectively. In CLM, the domain alignment of pixel-level feature and instance-level feature of SAR image objects is first implemented, and then to make the feature alignment of both domains more accurate, better aggregation of proposals of different classes of prototype objects in the same domain is required, and further compensation loss is proposed to further restrict the prototype alignment of SAR object features in both domains. We conduct experiments on two SAR datasets obtained from different satellites whose results show the superiority of the proposed method over the state-of-the-art (SOTA) methods and the effectiveness of the proposed module in improving the detection accuracy.

An unsupervised domain adaptation scheme for single-stage artwork recognition in cultural sites

Insulator defect detection is a crucial task in the intelligent inspection of transmission lines. Currently, there are challenges such as insufficient image samples and difficulties in annotation. Artificially generating samples is a feasible solution, but there are discrepancies between artificial and real sample distributions. This article proposes a cross-domain multilevel feature alignment R-CNN network. It uses a large number of artificially labeled images and a small number of unlabeled real images as the source and target domains, respectively. Based on Faster R-CNN, an instance-level feature adaptive alignment module is constructed. Different gradient adaptive training strategies are employed for the source and target domains to better achieve cross-domain instance-level feature alignment. An image-level multiscale local feature aggregation (MLA) module is built to achieve cross-domain image-level local feature alignment. A global feature alignment (GFA) module is also constructed to achieve cross-domain image-level global feature alignment. In the insulator defect detection experiment, the average precision (AP) at the Intersection over Union of 0.5 (AP50) of the proposed method is 6.8% higher than that of the baseline model, 8.6% higher than that of Faster-RCNN trained with artificial samples alone, and 1.1% higher than that of Faster-RCNN trained with real samples alone. Moreover, the proposed method achieves the highest accuracy on both self-built and public datasets, exhibiting an average improvement of 7.7% and 8.6% in AP50, respectively, compared to multiple comparable methods. The code for this article can be found at <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/n-tong/CMFAA-R-CNN</uri>.

Robust nuclei detection is crucial prerequisite for histologic characteristics of nuclei that can assist various clinical tasks such as disease diagnosis and cancer grading. Despite of their success, most existing nuclei detection methods ignore the case where the testing (target) domain has different data distribution with the training (source) domain, which is known as the problem of domain shift. In fact, the domain shift problem is prevalent in histopathology images due to various reasons such as different staining procedures and organ specific nuclear morphology. Thus, the performance of a nuclei detection model in the source domain will be hurt if it is directly applied to the target domain. To address this problem, we propose a novel instance-aware domain adaption framework for nuclei detection in histopathology images, which includes both image-level alignment (IMA) and instance-level alignment (INA) components to minimize the domain shift. Especially, INA component extracts instance-level features by using nuclei locations as the guidance and effectively aligns the instance-level features via adversarial training. Furthermore, to facilitate instance-level feature alignment, a Temporal Ensembling based Nuclei Localization (TENL) module is introduced in INA component to automatically generate candidate nuclei locations in the target domain. We evaluate the proposed method on different benchmark settings and obtain remarkable improvements compared to existing methods on the challenging problem of cross-domain cell nuclei detection.

We address the problem of federated domain generalization in an unsupervised setting for the first time. We first theoretically establish a connection between domain shift and alignment of gradients in unsupervised federated learning and show that aligning the gradients at both client and server levels can facilitate the generalization of the model to new (target) domains. Building on this insight, we propose a novel method named FedGaLA, which performs gradient alignment at the client level to encourage clients to learn domain-invariant features, as well as global gradient alignment at the server to obtain a more generalized aggregated model. To empirically evaluate our method, we perform various experiments on four commonly used multi-domain datasets, PACS, OfficeHome, DomainNet, and TerraInc. The results demonstrate the effectiveness of our method which outperforms comparable baselines. Ablation and sensitivity studies demonstrate the impact of different components and parameters in our approach.

Deep Joint Semantic Adaptation Network for Multi-source Unsupervised Domain Adaptation

Automatic and accurate segmentation of the ventricles and myocardium from multi-sequence cardiac MRI (CMR) is crucial for the diagnosis and treatment management for patients suffering from myocardial infarction (MI). However, due to the existence of domain shift among different modalities of datasets, the performance of deep neural networks drops significantly when the training and testing datasets are distinct. In this paper, we propose an unsupervised domain alignment method to explicitly alleviate the domain shifts among different modalities of CMR sequences, \emph{e.g.,} bSSFP, LGE, and T2-weighted. Our segmentation network is attention U-Net with pyramid pooling module, where multi-level feature space and output space adversarial learning are proposed to transfer discriminative domain knowledge across different datasets. Moreover, we further introduce a group-wise feature recalibration module to enforce the fine-grained semantic-level feature alignment that matching features from different networks but with the same class label. We evaluate our method on the multi-sequence cardiac MR Segmentation Challenge 2019 datasets, which contain three different modalities of MRI sequences. Extensive experimental results show that the proposed methods can obtain significant segmentation improvements compared with the baseline models.