Search Results (113)

Software maintenance is one of the most expensive phases in software development, especially when complex source code is the only available artifact. Clustering software modules and generating a structured architectural model can significantly reduce the effort and cost of maintenance. This study aims to achieve high-quality modularization by maximizing intra-cluster cohesion, minimizing inter-cluster coupling, and optimizing overall modular quality. Since finding optimal clustering is an NP-complete problem, many existing methods suffer from poor modular structures, instability, and inconsistent results. To overcome these limitations, this paper proposes a module clustering method using a discrete bedbug optimizer. In software architecture, symmetry refers to the balanced and structured arrangement of modules. In the proposed method, module clustering aims to identify and group related modules based on structural and behavioral similarities, reflecting symmetrical properties in the source code. Conversely, asymmetries, such as modules with irregular dependencies, can indicate architectural flaws. The method was evaluated on ten widely used real-world software datasets. The experimental results show that the proposed algorithm consistently delivers superior modularization quality, with an average score of 2.806 and a well-balanced trade-off between cohesion and coupling. Overall, this research presents an effective solution for software module clustering and provides better architecture recovery and more maintainable systems. Full article

Federated learning (FL) has emerged as a powerful framework for decentralized model training, preserving data privacy by keeping datasets localized on distributed devices. However, data heterogeneity, characterized by significant variations in size, statistical distribution, and composition across client datasets, presents a persistent challenge that impairs model performance, compromises generalization, and delays convergence. To address these issues, we propose FedDPA, a novel framework that utilizes dynamic prototypical alignment. FedDPA operates in three stages. First, it computes class-specific prototypes for each client to capture local data distributions, integrating them into an adaptive regularization mechanism. Next, a hierarchical aggregation strategy clusters and combines prototypes from similar clients, which reduces communication overhead and stabilizes model updates. Finally, a contrastive alignment process refines the global model by enforcing intra-class compactness and inter-class separation in the feature space. These mechanisms work in concert to mitigate client drift and enhance global model performance. We conducted extensive evaluations on standard classification benchmarks—EMNIST, FEMNIST, CIFAR-10, CIFAR-100, and Tiny-ImageNet 200—under various non-identically and independently distributed (non-IID) scenarios. The results demonstrate the superiority of FedDPA over state-of-the-art methods, including FedAvg, FedNH, and FedROD. Our findings highlight FedDPA’s enhanced effectiveness, stability, and adaptability, establishing it as a scalable and efficient solution to the critical problem of data heterogeneity in federated learning. Full article

To address issues commonly observed during the inference phase of large language models—such as inconsistent labels, formatting errors, or semantic deviations—a series of targeted strategies has been proposed. First, a relation label refinement strategy based on semantic similarity and syntactic structure has been designed to calibrate the model’s outputs, thereby improving the accuracy and consistency of label prediction. Second, to meet the contextual modeling needs of different types of instance bags, a multi-level contextual augmentation strategy has been constructed. For multi-sentence instance bags, a graph-based retrieval enhancement mechanism is introduced, which integrates intra-bag entity co-occurrence networks with document-level sentence association graphs to strengthen the model’s understanding of cross-sentence semantic relations. For single-sentence instance bags, a semantic expansion strategy based on term frequency-inverse document frequency is employed to retrieve similar sentences. This enriches the training context under the premise of semantic consistency, alleviating the problem of insufficient contextual information. Notably, the proposed multi-granularity framework captures semantic symmetry between entities and relations across different levels of context, which is crucial for accurate and balanced relation understanding. The proposed methodology offers practical advancements for semantic analysis applications, particularly in knowledge graph development. Full article

In recent years, cross-domain few-shot learning (CDFSL) has demonstrated remarkable performance in hyperspectral image classification (HSIC), partially alleviating the distribution shift problem. However, most domain adaptation methods rely on similarity metrics to establish cross-domain class matching, making it difficult to simultaneously account for intra-class sample size variations and inherent inter-class differences. To address this problem, existing studies have introduced a class weighting mechanism within the prototype network framework, determining class weights by calculating inter-sample similarity through distance metrics. However, this method suffers from a dual limitation: susceptibility to noise interference and insufficient capacity to capture global class variations, which may lead to distorted weight allocation and consequently result in alignment bias. To solve these issues, we propose a novel class-discrepancy dynamic weighting-based cross-domain FSL (CDDW-CFSL) framework. It integrates three key components: (1) the class-weighted domain adaptation (CWDA) method dynamically measures cross-domain distribution shifts using global class mean discrepancies. It employs discrepancy-sensitive weighting to strengthen the alignment of critical categories, enabling accurate domain adaptation while maintaining feature topology; (2) the class mean refinement (CMR) method incorporates class covariance distance to compute distribution discrepancies between support set samples and class prototypes, enabling the precise capture of cross-domain feature internal structures; (3) a novel multi-dimensional feature extractor that captures both local spatial details and continuous spectral characteristics simultaneously, facilitating deep cross-dimensional feature fusion. The results in three publicly available HSIC datasets show the effectiveness of the CDDW-CFSL. Full article

UAV-based small target detection is crucial in environmental monitoring, circuit detection, and related applications. However, UAV images often face challenges such as significant scale variation, dense small targets, high inter-class similarity, and intra-class diversity, which can lead to missed detections, thus reducing performance. To solve these problems, this study proposes a lightweight and high-precision model UAV-YOLO based on YOLOv8s. Firstly, a double separation convolution (DSC) module is designed to replace the Bottleneck structure in the C2f module with deep separable convolution and point-by-point convolution fusion, which can reduce the model parameters and calculation complexity while enhancing feature expression. Secondly, a new SPPL module is proposed, which combines spatial pyramid pooling rapid fusion (SPPF) with long-distance dependency modeling (LSKA) to improve the robustness of the model to multi-scale targets through cross-level feature association. Then, DyHead is used to replace the original detector head, and the discrimination ability of small targets in complex background is enhanced by adaptive weight allocation and cross-scale feature optimization fusion. Finally, the WIPIoU loss function is proposed, which integrates the advantages of Wise-IoU, MPDIoU and Inner-IoU, and incorporates the geometric center of bounding box, aspect ratio and overlap degree into a unified measure to improve the localization accuracy of small targets and accelerate the convergence. The experimental results on the VisDrone2019 dataset showed that compared to YOLOv8s, UAV-YOLO achieved an 8.9% improvement in the recall of mAP@0.5 and 6.8%, while the parameters and calculations were reduced by 23.4% and 40.7%, respectively. Additional evaluations of the DIOR, RSOD, and NWPU VHR-10 datasets demonstrate the generalization capability of the model. Full article

Background and Objectives: At present, intra-abdominal adhesions (IAAs) continue to be an important problem in surgery due to morbidity and mortality risks. Thymoquinone (TQ) and platelet-rich plasma (PRP) are molecules with known anti-inflammatory and antioxidant effects. However, a limited number of studies have investigated their efficacy in IAAs. In this study, we aimed to demonstrate the efficacy of TQ and PRP in reducing the development of IAAs and determine which molecule is more advantageous using an experimental animal model. Materials and Methods: Fifty-five male Wistar albino rats were included in the study. Five rats were used to obtain PRP, while fifty rats were randomly assigned to five groups (n = 10 per group): group I (sham) did not receive any treatment; group II (control) received no treatment after a cecum hemorrhage procedure; group III (saline) received 1 mL of saline treatment around the cecum after hemorrhage; group IV (PRP) received 1 mL of PRP (containing 3 × 10⁶ platelets/mL) around the cecum after hemorrhage; and group V (TQ) received 1 mL of TQ (containing 2 mg/mL TQ) around the cecum after hemorrhage. On the 10th day, IL1-β, TNF-α, E-selectin, and P-selectin levels were measured from the blood serum samples, and the cecum was histopathologically evaluated. Results: The lowest adhesion formation in terms of biochemical parameters was obtained in the TQ group (p < 0.05). Histopathological evaluations showed that saline, PRP, and TQ treatments were all effective, but none was superior. Conclusions: When histopathologically evaluated, saline, TQ, and PRP have similar effects in IAAs. However, when evaluated in terms of biochemical parameters, TQ prevented the formation of intra-abdominal adhesions more effectively than saline or PRP, owing to its strong anti-inflammatory and antioxidant properties. Full article

Aiming at the problems of existing time series data clustering methods, such as the lack of similarity metric universality, the influence of dimensional catastrophe, and the limitation of feature expression ability, a time series data clustering method based on unsupervised contrasting learning (UCL-TSC) is proposed. The method first utilizes Residual, TCN, and CNN-TCN to construct multi-view representations of spatial, temporal, and spatial–temporal features of time series data, and adaptively fuses complementary information to enhance feature extraction capabilities. Subsequently, positive and negative sample pairs are constructed based on nearest neighbor and pseudo-clustering label information. Finally, a contrast loss function consisting of feature loss, clustering loss, and a regularization term is designed to facilitate the model in achieving compact intra-cluster and sparse inter-cluster clustering effects in the clustering process. The experimental results on the UCR dataset show that UCL-TSC performs well with respect to several evaluation indexes, such as clustering accuracy, normalized information degree, and purity, and is more effective in learning time series data features and achieving accurate clustering compared to traditional clustering and deep clustering methods. Full article

In recent years, the global commercial aerospace industry has flourished, witnessing a rapid surge in customized satellite services. Deep learning has emerged as a pivotal tool for accurately identifying wetland vegetation. However, hyperspectral remote sensing images are often plagued by varying degrees of noise during acquisition, leading to subtle differences in spectral responses. Currently, vegetation classification models are tailored specifically for each hyperspectral sensor, making it challenging to generalize a model designed for one sensor to others. Furthermore, discrepancies in data distribution between training and test sets result in a notable decline in model performance, impeding model sharing across satellite hyperspectral sensors and hindering the interpretation of wetland scenes. Domain adaptation methods leveraging Generative Adversarial Networks (GANs) have been extensively researched and applied in the realm of cross-sensor land feature classification. Nevertheless, these data-level cross-domain classification strategies typically focus on band selection or alignment using relatively similar data to address image differences, without addressing spectral variability or incorporating pseudo-labels to enhance classification accuracy. Noise changes aggravate the distribution characteristics and model differences of vegetation in classification tasks. This has a negative impact on subsequent classification accuracy. To alleviate these problems, we have designed a linear unbiased stochastic network classification framework based on adversarial learning. The framework employs a style randomization algorithm to simulate spectral drift. It generates simulated images to enhance the model’s generalization ability. Supervised contrastive learning is utilized to prevent redundant learning of the same training images. Domain discrimination and domain-invariant characteristics are considered. We optimize the generator and discriminator using inter-class and intra-class contrast loss functions. The dual regularization training method is adopted, and non-redundant expansion is realized. It achieves similarity and addresses offsets. This method minimizes computational cost. Cross-sensor classification experiments were conducted, with comparative tests performed on a self-made wetland dataset. This method demonstrates significant advantages in wetland vegetation classification. According to the visualization results, our classification strategy can be used for cross-domain vegetation classification in coastal wetlands. It can also be applied to other small-satellite hyperspectral images and cross-satellite multispectral data, reducing on-site sampling costs and proving cost-effective. Full article

In data-scarcity scenarios, few-shot object detection (FSOD) methods exhibit a notable advantage in alleviating the over-fitting problem. Currently, research on FSOD in the field of remote sensing is advancing rapidly and FSOD methods based on the fine-tuning paradigm have initially displayed their excellent performance. However, existing fine-tuning methods often encounter classification confusion issues. This is potentially because of the shortage of explicit modeling for transferable common knowledge and the biased class distribution, especially for fine-grained targets with higher inter-class similarity and intra-class variance. In view of this, we first propose a decoupled self-distillation (DSD) method to construct class prototypes in two decoupled feature spaces and measure inter-class correlations as soft labels or aggregation weights. To ensure a robust set of class prototypes during the self-distillation process, we devise a feature filtering module (FFM) to preselect high-quality class representative features. Furthermore, we introduce a progressive prototype calibration module (PPCM) with two steps, compensating the base prototypes with the prior base distribution and then calibrating the novel prototypes with adjacent calibrated base prototypes. Experiments on MAR20 and customized SHIP20 datasets have demonstrated the superior performance of our method compared to other existing advanced FSOD methods, simultaneously confirming the effectiveness of all proposed components. Full article

Fine-grained image classification is faced with the challenge of significant intra-class differences and subtle similarities between classes, with a limited number of labelled data. Previous few-shot learning approaches, however, often fail to recognize these discriminative details, such as a bird’s eyes and beak. In this paper, we proposed a channel-wise attention-enhanced feature mutual reconstruction mechanism that helps to alleviate these problems for fine-grained image classification. This mechanism first employed a channel-wise attention module (CAM) to learn the channel weights for both the support and query features. We utilized channel-wise self-attention to assign greater importance to object-relevant channels. This helps the model to focus on subtle yet discriminative details, which is essential to the classification process. Then, we introduce a feature mutual reconstruction module (FMRM) to reconstruct features. The support features are reconstructed by a support-weight-enhanced feature map to reduce the intra-class variations, and query features are reconstructed by a query-weight-enhanced feature map to increase inter-class variations. The results of classification depend on the similarity between reconstructed features and enhanced features. We evaluated the performance based on four fine-grained image datasets when Conv-4 and Resnet-12 were used. The experimental results showed that our method outperforms previous few-shot fine-grained classification methods. This proves that our method can improve fine-grained image classification performance and simultaneously balance both the inter-class and intra-class variations. Full article

There are many evolving dynamic networks in the real world, and community detection in dynamic networks is crucial in many complex network analysis applications. In this paper, a consensus community-based discrete spider wasp optimization (SWO) approach is proposed for the dynamic network community detection problem. First, the coding, initialization, and updating strategies of the spider wasp optimization algorithm are discretized to adapt to the community detection problem. Second, the concept of intra-population and inter-population consensus community is proposed. Consensus community is the knowledge formed by the swarm summarizing the current state as well as the past history. By maintaining certain inter-population consensus community during the evolutionary process, the population in the current time window can evolve in a similar direction to those in the previous time step. Experimental results on many artificial and real dynamic networks show that the proposed method produces more accurate and robust results than current methods. Full article

It is a great challenge for a safe surgery to localize the cutting tip during laminar grinding. To address this problem, we develop a framework of state estimation based on the CT image–force model. For the proposed framework, the pre-operative CT image and intra-operative milling force signal work as source inputs. In the framework, a bone milling force prediction model is built, and the surgical planned paths can be transformed into the prediction sequences of milling force. The intra-operative milling force signal is segmented by the tumbling window algorithm. Then, the similarity between the prediction sequences and the segmented milling signal is derived by the dynamic time warping (DTW) algorithm. The derived similarity indicates the position of the cutting tip. Finally, to overcome influences of some factors, we used the random sample consensus (RANSAC). The code of the functional simulations has be opened. Full article

The rapid growth of the global population, the acceleration of the urbanization process, and the demands of economic development, place enormous pressure on scarce land resources. Cultivated land use presents a series of problems, hindering its socioeconomic and ecological sustainability. The sustainable intensification of cultivated land use (SICLU) is a development model designed to maximize land use efficiency, while minimizing environmental pollution. It is considered to be an efficient method to achieve three aspects of sustainable goals, namely in regard to society, the economy, and ecology, simultaneously. This approach has significant theoretical and practical implications for China’s food security and ecological safety. This study incorporates the “agricultural carbon emissions” indicator into the indicator evaluation system. Using the super-efficiency SBM model, we estimate the SICLU levels in China from 2001 to 2020. ArcGIS and the Dagum Gini coefficient decomposition model are employed to explore the temporal and spatial evolution characteristics and non-equilibrium spatial dynamics of SICLU in China. Finally, the Tobit regression model is used to reveal the driving factors. The results show the following: (1) Since 2003, China’s SICLU levels demonstrate an overall ascent amid fluctuations, sustaining a relatively high average annual level of 0.945. (2) In terms of spatial evolution patterns, China’s SICLU levels demonstrate significant spatial disparities, with distinct differences among the four major regions. Regions with similar SICLU levels show a certain degree of spatial adjacency. (3) There are significant regional disparities in China’s SICLU levels, which overall exhibit a declining trend. The differences between regions are the primary source of spatial variation, followed by hypervariable density and intra-regional disparities. (4) The regional industrial structure, the level of agricultural modernization, the agricultural cropping structure, and the per capita sown area, positively influence the enhancement of SICLU levels in China. Throughout the study period, the SICLU levels in China continuously improved and the overall regional disparities diminished. However, significant inter-regional imbalances persist, necessitating tailored optimization measures, based on local conditions. Establishing a coordinated mechanism for orderly and synergistic regional development is crucial, in order to provide references to decision-makers to promote the rational use of arable land in China. Full article

Few-shot fine-grained image classification (FSFGIC) aims to classify subspecies with similar appearances under conditions of very limited data. In this paper, we observe an interesting phenomenon: different types of image data augmentation techniques have varying effects on the performance of FSFGIC methods. This indicates that there may be biases in the features extracted from the input images. The bias of the acquired feature may cause deviation in the calculation of similarity, which is particularly detrimental to FSFGIC tasks characterized by low inter-class variation and high intra-class variation, thus affecting the classification accuracy. To address the problems mentioned, we propose an unbiased feature estimation network. The designed network has the capability to significantly optimize the quality of the obtained feature representations and effectively reduce the feature bias from input images. Furthermore, our proposed architecture can be easily integrated into any contextual training mechanism. Extensive experiments on the FSFGIC tasks demonstrate the effectiveness of the proposed algorithm, showing a notable improvement in classification accuracy. Full article

Porcine respiratory disease is a significant economic problem for the global swine industry. Haemophilus parasuis (H. parasuis), Streptococcus suis (S. suis), and Pasteurella multocida (P. multocida) are three important pathogenic bacteria of the swine respiratory tract. Notably, the three pathogens not only frequently manifest as mixed infections, but their striking clinical similarities also present difficulties for pig populations in terms of disease prevention and treatment. Thus, we developed a triplex real-time quantitative polymerase chain reaction (qPCR) assay based on a TaqMan probe for the detection of H. parasuis, S. suis serotype 2, and P. multocida. Primers and probes were designed to target the conserved regions of the H. parasuis OmpP2 gene, the S. suis serotype 2 gdh gene, and the P. multocida Kmt1 gene. By optimizing the reaction system and conditions, a triplex qPCR method for simultaneous detection of H. parasuis, S. suis serotype 2, and P. multocida was successfully established. The amplification efficiencies of the standard curves for all three pathogens were found to be highly similar, with values of 102.105% for H. parasuis, 105.297% for S. suis serotype 2, and 104.829% for P. multocida, and all R² values achieving 0.999. The specificity analysis results showed that the triplex qPCR method had a strong specificity. The sensitivity test results indicated that the limit of detection can reach 50 copies/μL for all three pathogens. Both intra- and inter-assay coefficients of variation for repeatability were below 1%. This triplex qPCR method was shown to have good specificity, sensitivity, and reproducibility. Finally, the triplex qPCR method established in this study was compared with the nested PCR as recommended by the Chinese national standard (GB/T34750-2017) for H. parasuis, the PCR as recommended by the Chinese national standard (GB/T 19915.9-2005) for S. suis serotype 2, and the PCR as recommended by the Chinese agricultural industry standard (NY/T 564-2016) for P. multocida by detecting the same clinical samples. Both methods are reasonably consistent, while the triplex qPCR assay was more sensitive. In summary, triplex qPCR serves not only as a rapid and accurate detection and early prevention method for these pathogens but also constitutes a robust tool for microbial quality control in specific pathogen-free pigs. Full article