Applied Intelligence最新文献

{"title":"RDM2: a two-stage model based on residual learning diffusion model and multi-scale convolution for Low Dose CT denoising","authors":"Zhencun Jiang,&nbsp;Kangrui Ren,&nbsp;Kefan Wang,&nbsp;Zhongjie Wang","doi":"10.1007/s10489-025-06604-0","DOIUrl":"10.1007/s10489-025-06604-0","url":null,"abstract":"<div><p>Computed Tomography (CT) is widely used in clinical diagnosis, but large amount of radiation accompanied is not expected. Low Dose CT (LDCT) can reduce the radiation effect, however, noise and artifacts will be unavoidably produced. Low dose accompanies large noise intensity, which is difficult to effectively denoise while retaining the details. Aiming at this problem, a two-stage LDCT denoising model, named RDM2, is proposed. In the first stage, a residual learning diffusion model is constructed to eliminate the noise of LDCT. The residuals between LDCT and Normal Dose CT (NDCT) is a kind of complex mixed noise with unknown intensity. In order to fully utilize the residual information, the whole residual is equally divided into small pieces and added iteratively in the diffusion process. Considering even the best trained residual diffusion model may bring unavoidable error when it is used for prediction, a multi-scale convolution encoder decoder convolution neural network (MEDCNN) is proposed in the second stage to further reduce this part of error. The proposed model RDM2 is validated on both the Mayo2020 25% dose LDCT dataset and Mayo2020 10% dose LDCT dataset, the values of PSNR, SSIM, and RMSE on these two datasets are respectively 44.7651, 0.9939, 0.0068 and 35.5302, 0.9601, 0.0172. It is proved that RDM2 outperforms the traditional method, the supervised learning-based method and the GAN-based method, and has the potential to meet clinical needs. Code is available at: https://github.com/zhencunjiang/RDM2.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 13","pages":""},"PeriodicalIF":3.5,"publicationDate":"2025-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144904937","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Survey on Tabular Data Privacy and Synthetic Data Generation in Industry 4.0","authors":"Hadi Koubeissy,&nbsp;Amir Amine,&nbsp;Marc Kamradt,&nbsp;Abdallah Makhoul","doi":"10.1007/s10489-025-06823-5","DOIUrl":"10.1007/s10489-025-06823-5","url":null,"abstract":"<div><p>Synthetic data is an emerging field that solves the raised need for privacy-preserving data sharing and the lack of real data. One of the most common data types used is tabular data, which is widely used to train machine learning models, especially in the industrial domain for better decision-making and edge case handling, two key points in Industry 4.0. In this paper, we present and evaluate state-of-the-art models for tabular data generation under a proposed taxonomy consisting of statistical models, generative adversarial networks (GANs)-based models, denoising diffusion probabilistic models (DDPMs), and large language models (LLMs). Additionally, we propose a revised evaluation taxonomy consisting of three dimensions, including realism, representativeness, and privacy. The results proved that analyzing models based on multiple metrics from each category could ensure a better understanding of the dataset when used for downstream tasks. Finally, we found that models based on GANs are still a solid option in multiple cases, such as a constrained computational environment. In contrast, models based on LLMs and DDPMs are more promising in terms of realism and representativeness. More research should be invested in overcoming limitations such as numerical data representation and long training times for LLMs. Our survey serves as a study for existing models and newer directions in the field, with guidelines for evaluation that can be applied to industrial and other domains.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 13","pages":""},"PeriodicalIF":3.5,"publicationDate":"2025-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144904974","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Federated trajectory clustering based on multi-feature similarity calculation","authors":"Kun Guo,&nbsp;Xinglong Hu,&nbsp;Zhiyu Zhang,&nbsp;Chuyu Liu,&nbsp;Qishan Zhang","doi":"10.1007/s10489-025-06813-7","DOIUrl":"10.1007/s10489-025-06813-7","url":null,"abstract":"<div><p>Trajectory clustering plays an important role in numerous real-world applications, such as urban transportation planning and tourist route recommendation. Existing trajectory clustering approaches primarily focus on the spatial and temporal features of trajectories but neglect the velocity feature. Therefore, it is difficult for them to distinguish trajectories sharing spatial and temporal features but diverging velocities. Furthermore, in the context of distributed trajectory clustering among multiple participants, individuals’ privacy, such as the travel routes or habits of a person, should never be violated, which necessitates the equipment of trajectory clustering with privacy-preserving techniques. In this paper, we propose a Federated and Multi-Feature-based Trajectory Clustering (FMFTC) algorithm to address the above issues. First, we develop a Multi-Feature-based Trajectory Clustering (MFTC) algorithm with a new multi-feature to vector encoder (MF2Vec) to capture spatial, temporal and velocity features during trajectory embedding generation. Second, we adapt MFTC to the federated learning paradigm to construct FMFTC for privacy-preserving distributed trajectory clustering. The experiments on real-world datasets demonstrate that FMFTC achieves up to <span>(varvec{24.4%})</span> higher accuracy than existing trajectory clustering algorithms and performs identically as MFTC with no accuracy loss.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 13","pages":""},"PeriodicalIF":3.5,"publicationDate":"2025-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144904976","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multi-agent neighborhood coordinated and holistic optimized actor-critic framework for adaptive traffic signal control","authors":"Qi Deng,&nbsp;Lijun Wu,&nbsp;Zhiyuan Li,&nbsp;Kaile Su,&nbsp;Wei Wu,&nbsp;Weiwei Duan","doi":"10.1007/s10489-025-06758-x","DOIUrl":"10.1007/s10489-025-06758-x","url":null,"abstract":"<div><p>Adaptive Traffic Signal Control (ATSC) is a pivotal research area within intelligent transportation systems, aiming to enhance transportation efficiency and alleviate traffic congestion at signalized intersections. While multi-agent deep reinforcement learning has been extensively applied to ATSC, existing approaches commonly frame it as a fully cooperative problem, presupposing that all agents are committed to pursuing a collective optimal solution. However, achieving such altruistic cooperation is often impractical. Furthermore, as the number of agents escalates, challenges such as the curse of dimensionality and non-stationarity arise, complicating the learning process. To address these issues, we propose a novel perspective by framing ATSC as a competitive-cooperative game trade-off scenario and design a multi-agent framework, termed Neighborhood Coordinated and Holistic Optimized Actor-Critic (NcHo-AC). Specifically, we introduce a novel traffic state representation, design a sophisticated feature extraction network, develop a robust training algorithm, and leverage mean field approximation to model population-level agent interactions. These designs foster neighborhood-level cooperation and communication, facilitate the learning of the desired Nash equilibrium, and mitigate the noise caused by agents’ exploratory behaviors, thereby alleviating non-stationarity and the curse of dimensionality, while enhancing scalability to large-scale traffic networks. Comprehensive experiments conducted on both synthetic and real-world datasets demonstrate that NcHo-AC significantly outperforms state-of-the-art baselines across four key metrics: average travel time, average queue length, delay, and throughput, along with improved convergence, robustness, and interpretability.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 13","pages":""},"PeriodicalIF":3.5,"publicationDate":"2025-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144904977","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Uncovering connections: a reference network approach to statute law retrieval","authors":"Thi-Hai-Yen Vuong,&nbsp;Hai-Long Nguyen,&nbsp;Tan-Minh Nguyen,&nbsp;Ha-Thanh Nguyen,&nbsp;Le-Minh Nguyen,&nbsp;Xuan-Hieu Phan","doi":"10.1007/s10489-025-06818-2","DOIUrl":"10.1007/s10489-025-06818-2","url":null,"abstract":"<div><p>The increasing volume and complexity of statute law data have led to a growing demand for efficient and effective retrieval methods. This paper presents a novel approach to statute law retrieval that utilizes reference networks to uncover connections between laws. By representing law articles as a network of references, our method allows users to quickly identify relevant direct and indirect articles. The key point is that the reference network can encode both internal and external legal relations, helping to integrate both the local and the long-range dependencies into the final retrieval model. The proposed approach is evaluated on several statute law corpora and shows that it performs better existing methods on the same tasks. In addition, our finding is that internal references help enhance the accuracy significantly while external links are also important. Our empirical study also suggests the optimal range of local window size to achieve a balance between retrieval accuracy and noise. Our approach can also contribute to the development of AI-assisted legal research tools, making it easier for legal practitioners to find relevant laws and precedents. Furthermore, by uncovering hidden connections between laws, our method can help identify inconsistencies and gaps in the legal system, ultimately improving its effectiveness and reliability.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 13","pages":""},"PeriodicalIF":3.5,"publicationDate":"2025-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144904975","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Simultaneous fault prediction in evolving industrial environments with ensembles of Hoeffding adaptive trees","authors":"A. Esteban,&nbsp;A. Cano,&nbsp;S. Ventura,&nbsp;A. Zafra","doi":"10.1007/s10489-025-06786-7","DOIUrl":"10.1007/s10489-025-06786-7","url":null,"abstract":"<div><p>Predictive Maintenance (PdM) emerges as a critical task of Industry 4.0, driving operational efficiency, minimizing downtime, and reducing maintenance costs. However, real-world industrial environments present unsolved challenges, especially in predicting simultaneous and correlated faults under evolving conditions. Traditional batch-based and deep learning approaches for simultaneous fault prediction often fall short due to their assumptions of static data distributions and high computational demands, making them unsuitable for dynamic, resource-constrained systems. In response, we propose OEMLHAT (Online Ensemble of Multi-Label Hoeffding Adaptive Trees), a novel model tailored for real-time, multi-label fault prediction in non-stationary industrial settings. OEMLHAT introduces a scalable online ensemble architecture that integrates online bagging, dynamic feature subspacing, and adaptive output weighting. This design allows it to efficiently handle concept drift, high-dimensional input spaces, and label sparsity, key bottlenecks in existing PdM solutions. Experimental results on three public multi-label PdM case studies demonstrate substantial improvements in predictive performance of OEMLHAT over previous batch-based and online proposals for multi-label classification, particularly with an average improvement in micro-averaged F1-score of 18.49% over the second most-accurate batch-based proposal and of 8.56% in the case of the second best online model. By addressing a critical gap in online multi-label learning for PdM, this work provides a robust and interpretable solution for next-generation industrial monitoring systems for fault detection, particularly for rare and concurrent failures.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 13","pages":""},"PeriodicalIF":3.5,"publicationDate":"2025-08-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s10489-025-06786-7.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144892411","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Remote sensing image change detection method based on dual-branch multi-level feature difference interactive learning","authors":"Songtao Ding,&nbsp;Xinyu Li,&nbsp;Hongyu Wang,&nbsp;Shiwen Gao","doi":"10.1007/s10489-025-06728-3","DOIUrl":"10.1007/s10489-025-06728-3","url":null,"abstract":"<div><p>Remote sensing (RS) image change detection (CD) is a key technology in environmental monitoring and geographic information systems (GIS). It can reveal the dynamic changes of surface features and is of great significance in fields such as urban planning, disaster assessment, and ecological research. However, the pseudo-change problem, that is, the image differences caused by non-actual surface changes, often affects the accuracy of detection, leading to false alarms and omissions, which limits the effectiveness of the CD technology. Traditional dual-branch CD methods often focus on basic feature extraction. This method independently processes the feature extraction of the bi-temporal phases and lacks a comparative interactive learning process for the features of the bi-temporal phases, thereby weakening its ability to identify pseudo-changes in complex environments. To solve the above problems, we propose a RS image CD method based on dual-branch multi-level feature difference interactive learning (DMFDIL). The model is built based on the siamese convolutional neural network (CNN) of deep learning and includes three parts: the dual-branch cooperative coding module (DCM), the dual-branch difference decoding module (DDDM), and the change output module (COM). Among them, the DCM innovatively introduces the tri-attention mechanism. Through this mechanism, the model can effectively interact on multi-level features, enhancing the ability to capture subtle changes in RS images, especially in distinguishing real changes from pseudo-changes. The DDDM, on the other hand, focuses on further optimizing the detection capability of the model by identifying real changes from pseudo-changes and integrating feature information at different scales. Finally, the validation was carried out on three public datasets, and the results were better than other popular methods. The experimental results on the LEVIR-CD dataset show that the proposed DMFDIL model achieved 95.80% in precision (Pre), 94.54% in recall (Rec), 95.16% in F1-score (F1), 91.10% in Intersection over Union (IoU), and 99.07% in overall accuracy (OA), which are significantly better than those of the state-of-the-art (SOTA) approaches. This method provides a new technical approach in the field of RS image CD, especially in improving detection accuracy and dealing with pseudo-change problems, and has important application value and broad application prospects.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 13","pages":""},"PeriodicalIF":3.5,"publicationDate":"2025-08-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144892410","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A non-adaptive segmentation algorithm for particle images in controlled environments with uniform backgrounds based on two-round superpixel segmentation and ensemble learning","authors":"Shiming Zhang,&nbsp;Zhikang Ma,&nbsp;Yan Ma","doi":"10.1007/s10489-025-06792-9","DOIUrl":"10.1007/s10489-025-06792-9","url":null,"abstract":"<div><p>Particle image segmentation under controlled environments with uniform backgrounds remains a challenging task due to issues such as particle adhesion, low contrast, and uneven illumination. Existing methods often suffer from over-segmentation or under-segmentation, especially when applied to microscopic or industrial particles. To address these problems, this paper proposes a non-adaptive segmentation algorithm called TS-EL (Two-round Superpixel Segmentation and Ensemble Learning), which is specifically designed for particle images captured in controlled settings with homogeneous backgrounds. The TS-EL framework performs coarse-to-fine superpixel segmentation and easy-to-hard classification. It introduces a gradient distance-based superpixel segmentation algorithm (GradSE) to improve boundary alignment between superpixels and particle contours. A Gaussian model and dual-factor classification criteria are employed to categorize high-confidence superpixels into foreground and background, while low-confidence regions are refined using a second-round segmentation based on minimum bounding boxes. The final classification of ambiguous regions is achieved via the LogitBoost ensemble learning algorithm. Experimental results on three types of particle images (grain, color masterbatch, and cell images) demonstrate that the proposed method outperforms seven state-of-the-art comparative algorithms in terms of segmentation accuracy and boundary adherence. The method is non-adaptive and relies on empirically set parameters, making it well-suited for batch processing in controlled environments but less generalizable to natural or complex scenes.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 13","pages":""},"PeriodicalIF":3.5,"publicationDate":"2025-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144891508","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"3D ME-Net: multi-scale and edge-guided enhancement network for intracranial aneurysm segmentation","authors":"Jiaqi Wang,&nbsp;Juntong Liu,&nbsp;Jun Li,&nbsp;Aiping Wu,&nbsp;Yunfeng Zhou,&nbsp;Mingquan Ye","doi":"10.1007/s10489-025-06779-6","DOIUrl":"10.1007/s10489-025-06779-6","url":null,"abstract":"<div><p>Intracranial aneurysms are relatively common and life-threatening conditions, making precise segmentation during early diagnosis crucial. However, the challenges of poor imaging quality and high noise levels often result in unclear aneurysm edges. Additionally, the varying sizes of aneurysms further complicate accurate segmentation. To address these issues, we propose a <b>M</b>ultiscale and <b>E</b>dge-guided enhanced 3D deep learning model. First, the asymmetrically larger network with enhanced hierarchical feature representation effectively captures subtle image features, thereby improving the localization of anatomical structures. Second, the multi-scale feature fusion mechanism within the encoder improves feature diversity and edge information, enhancing segmentation precision for aneurysms of different sizes. Finally, the edge-guided attention technique within the decoder combines local features with predicted heatmaps to extract comprehensive edge information. The experimental results demonstrate that the model outperforms general models in five key metrics on the internal dataset. External dataset testing confirms its adaptability and robustness across data from different acquisition protocols and hardware configurations. Clinical trials have further validated its practicality, assisting radiologists in more accurate intracranial aneurysm diagnosis.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 13","pages":""},"PeriodicalIF":3.5,"publicationDate":"2025-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144891509","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Memristive Bi-Path Wavelet Transformer for low-light image enhancement","authors":"Dirui Xie,&nbsp;Qi Cheng,&nbsp;Yue Zhou,&nbsp;Xiaofang Hu","doi":"10.1007/s10489-025-06771-0","DOIUrl":"10.1007/s10489-025-06771-0","url":null,"abstract":"<div><p>Images captured under low-light conditions are characterized by poor quality and insufficient exposure, which adversely affects the performance of downstream tasks, such as autonomous driving and nighttime surveillance. Recently, Transformer-based methods have achieved notable success in low-light image enhancement. However, these methods exhibit limited local information modeling capabilities and encounter issues with outliers due to insufficient dynamic range, which curtail their performance in low-light image enhancement. Additionally, the quadratic computational complexity of their Softmax-based self-attention mechanisms renders these methods challenging to deploy on edge devices. To address these issues, we propose a memristor-based Bi-Path Wavelet Transformer (BWT) with linear computational complexity. Specifically, we design a novel Dual-path Wavelet Linear Attention (BWLA) to replace the Softmax-based self-attention, enabling efficient local and global information extraction and aggregation at linear complexity. We propose a hardware implementation scheme of BWT based on memristors, which reduces deployment complexity and offers an effective solution for deploying low-light enhancement algorithms on edge devices. Experiments on multiple low-light enhancement benchmark datasets demonstrate that our method outperforms multiple state-of-the-art (SOTA) methods.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 13","pages":""},"PeriodicalIF":3.5,"publicationDate":"2025-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144868746","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}