Expert Systems with Applications最新文献

{"title":"Nighttime airport runway FOD intrusive detection through frequency-domain interference of spatially aggregated dynamic feature","authors":"Guangchen Chen,&nbsp;Yinhui Zhang,&nbsp;Zifen He,&nbsp;Ying Huang","doi":"10.1016/j.eswa.2025.128719","DOIUrl":"10.1016/j.eswa.2025.128719","url":null,"abstract":"<div><div>Accidental intrusion of foreign object debris (FOD) on the airport runway often causes fatal safety hazards in aviation transportation during take-off and landing especially at night as the degraded imaging quality. To address the false and missed detections in low-light images by current advanced methods, the Frequency-Domain Interference Network of spatially aggregated dynamic feature (FDI-Net) is proposed to improve the recognition accuracy of threatening FOD at nighttime. Firstly, to deal with the challenges posed by degraded low-quality imaging at night, we propose the Frequency-Domain Adaptive Tuning (FDAT) spatial pooling module, which utilizes fast Fourier transformation to construct frequency-domain features from the row and column pixels of FOD images. Subsequently, generating the wave function signal through the superposition of row and column components, and adaptively tuning the frequency and phase spectra using dynamically learnable weights within the network structure. This process effectively suppresses redundant information while enhancing the grayscale and texture feature space. Secondly, the Dynamic Granular Aggregated Interference (DGAI) module is developed to transform the FOD spatial features into amplitude and phase representations, enabling the extraction of fine-grained feature information through dynamic depth fusion. This module then aggregates the fine-grained features using the interference effects of sine and cosine waves to enhance positive FOD target information and suppress negative interference. Finally, the detection model is deployed on an embedded edge computing platform to develop a mobile nighttime airport runway foreign object debris detection system. Experimental results demonstrate that the proposed model effectively detects ten categories of small and medium-scale FOD targets, achieving optimal accuracy results of 97.9 %, 89.8 %, and 77.6 % on the mAP50, mAP75, and mAP50-95 metrics, respectively. In addition, our method achieves the accuracy of 94.5 %, 89.9 %, 85.2 %, 76.2 %, 92.4 %, 92.2 %, 95.5 %, 74.4 %, 97.5 %, and 99.5 % on ten categories, respectively. Consequently, the intelligent detection system holds significant potential for preventing accidents caused by FOD in the aviation transportation safety field.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"294 ","pages":"Article 128719"},"PeriodicalIF":7.5,"publicationDate":"2025-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144518227","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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":"EfficientPEAL: Efficient prior-embedded attention learning for partially overlapping point cloud registration","authors":"Junle Yu ,&nbsp;Wenhui Zhou ,&nbsp;Zhehao Shen ,&nbsp;Yongwei Miao","doi":"10.1016/j.eswa.2025.128591","DOIUrl":"10.1016/j.eswa.2025.128591","url":null,"abstract":"<div><div>Learning discriminative point-wise features is critical for partially overlapping point cloud registration. In recent years, the integration of a Transformer into point cloud feature representation has demonstrated remarkable success, which typically involves a self-attention module to learn intra-point-cloud features, followed by a cross-attention module for feature exchange between input point clouds. Transformer models mainly benefit from the use of self-attention to capture the global correlations in feature space. However, the global correlations involved in self-attention may not only result in a significant amount of redundant computational overhead but also introduce feature ambiguities, especially in low-overlap scenarios. This is because overlapping regions of point clouds typically do not span a wide range but are rather concentrated around a localized area. Therefore, the correlations with an extensive range of non-overlapping points are ineffective and may degrade the discriminability of features. To address this issue, we present a <strong>E</strong>fficient <strong>P</strong>rior-<strong>E</strong>mbedded <strong>A</strong>ttention <strong>L</strong>earning model (<strong>E</strong>fficientPEAL). By incorporating overlap prior to the learning process, the point clouds are divided into two parts. One part includes points lying in the putative overlapping region and the other includes points located in the putative non-overlapping region. Then, EfficientPEAL performs localized attention with the putative overlapping points. The proposed attention module significantly reduces the computational complexity of the model while achieving competitive performance. Extensive experiments on 3DMatch/3DLoMatch, ScanNet, and KITTI datasets demonstrate its effectiveness.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"293 ","pages":"Article 128591"},"PeriodicalIF":7.5,"publicationDate":"2025-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144481021","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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":"Multilingual neural machine translation by cascading computational graphs","authors":"Abouzar Qorbani ,&nbsp;Reza Ramezani ,&nbsp;Ahmad Baraani ,&nbsp;Arefeh Kazemi","doi":"10.1016/j.eswa.2025.128722","DOIUrl":"10.1016/j.eswa.2025.128722","url":null,"abstract":"<div><div>In the era of artificial intelligence, multilingual models have become increasingly vital in machine translation tasks. However, Multilingual Neural Machine Translation (MNMT) faces persistent challenges, notably reduced translation quality and language interference. When training on diverse language pairs, the translation performance for certain languages may degrade due to negative transfer effects. To address this problem, researchers have proposed various strategies such as parameter sharing, partial sharing, and language-specific parameterization. Despite these efforts, limitations remain—including high data requirements, reliance on linguistic relatedness, inflexibility in model architecture adaptation during training, and negative inference (producing output in an unintended language). The identification and targeted modification of effective and ineffective nodes within a neural model can effectively enhance the translation performance, particularly for low-resource and extremely low-resource languages. In this paper, a novel method is proposed in this study that identifies ineffective nodes in an MNMT model and corrects them by twinning with effective counterparts. This is achieved through computational graph grouping based on semantic similarity. The proposed method has been evaluated on several multilingual datasets, including TED2013, TED2020, and BIBLE. Relative to baseline models, the proposed method demonstrates notable improvements in BLEU scores—achieving relative gains of 23.7 % on TED2013, 7.06 % on TED2020, and 16.9 % on BIBLE. It also outperforms large-scale systems such as ChatGPT, Bing GPT-4, and Google Neural Machine Translation (GNMT) across all evaluated datasets. Furthermore, the performance has been assessed on the extremely low-resource language pair English–Igbo using the OPUS-100 dataset. The results show that the proposed method outperforms baseline models by 2.58 %, while the large-scale Madlad400-3B model, despite its depth (32 layers, 450 languages), struggles in this setting. Similarly, the Semlin-MNMT model performs well for high-resource pairs but shows significant degradation on low-resource languages. Overall, our proposed method provides a robust and scalable approach for enhancing MNMT quality in both one-to-many and many-to-many translation scenarios. Its effectiveness in low-resource and extremely low-resource settings highlights its practical value and contribution to the advancement of multilingual translation systems.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"294 ","pages":"Article 128722"},"PeriodicalIF":7.5,"publicationDate":"2025-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144510836","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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":"ChatSOS: Vector database augmented generative question answering assistant in safety engineering","authors":"Haiyang Tang,&nbsp;Dongping Chen,&nbsp;Qingzhao Chu,&nbsp;Zhenyi Liu","doi":"10.1016/j.eswa.2025.128644","DOIUrl":"10.1016/j.eswa.2025.128644","url":null,"abstract":"<div><div>With the rapid advancement of natural language processing technologies, generative artificial intelligence techniques, represented by large language models (LLMs), are gaining increasing prominence and demonstrating significant potential for applications in safety engineering. However, foundational LLMs face constraints such as limited training data coverage and unreliable responses. This study develops a vector database from 117 explosion accident reports in China spanning 2013 to 2023, employing techniques such as corpus segmenting and vector embedding. By utilizing the vector database, which outperforms the relational database in information retrieval quality, we provide LLMs with richer, more relevant knowledge. Comparative analysis of LLMs demonstrates that ChatSOS significantly enhances reliability, accuracy, and comprehensiveness, improves adaptability and clarification of responses. These results illustrate the effectiveness of supplementing LLMs with an external database, highlighting their potential to handle professional queries in safety engineering and laying a foundation for broader applications.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"294 ","pages":"Article 128644"},"PeriodicalIF":7.5,"publicationDate":"2025-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144510838","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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 synthetic data-enhanced method for automated 3D pose recognition of construction workers","authors":"Yonglin Fu ,&nbsp;Weisheng Lu ,&nbsp;Zhiming Dong ,&nbsp;Yihai Fang","doi":"10.1016/j.eswa.2025.128768","DOIUrl":"10.1016/j.eswa.2025.128768","url":null,"abstract":"<div><div>Automated 3D pose recognition of construction workers is instrumental to analyzing their occupational safety and health, productivity and other jobsite behaviors. Existing studies in this field have been confined to high-quality training datasets collected from real-life construction jobsites, potentially triggering ethical, privacy, and cost concerns. Inspired by the success of synthetic data in other fields, this research proposes a synthetic data-enhanced method for automated 3D pose recognition of construction workers. It generates a synthetic dataset to supplement a real-life dataset for model training, presents a monocular vision-based model for recognizing multiple workers’ 3D poses, and then validates the model performance. Experiments verify that this model jointly trained with synthetic and real data outperforms a model trained on real data alone. The data enrichment approach explored in this study offers reliable data quality at less expense than real data-focused approaches. This research therefore lays a foundation for a series of studies to enhance workers’ occupational safety and health and productivity.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"294 ","pages":"Article 128768"},"PeriodicalIF":7.5,"publicationDate":"2025-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144501576","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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 semi-supervised learning via globally guided pseudo-labeling: A robust approach for label-scarce scenarios","authors":"Yuan Xi,&nbsp;Qiong Li,&nbsp;Haokun Mao","doi":"10.1016/j.eswa.2025.128667","DOIUrl":"10.1016/j.eswa.2025.128667","url":null,"abstract":"<div><div>Federated Semi-Supervised Learning (FSSL) is a powerful paradigm for collaboratively training models on both labeled and unlabeled datasets, which is adopted in domains such as healthcare and IoT. However, heterogeneous data distributions and imbalanced labeling capabilities both lead to significant prediction bias across participating clients, further resulting in skewed pseudo-labels during the local training stage. Most existing FSSL studies address the bias by improving model consistency, which relies on a well-trained benchmark derived from the fully labeled client, and encounters challenges in label-scarce scenarios. In this paper, we propose a novel FSSL method, namely Federated Globally Guided pseudo-labeling (FedGGp), suitable for both label-scarce and Non-Independent and Identically Distributed (Non-IID) scenarios. Specifically, this strategy summarizes the prediction bias assessments based on skewed class predictions, and modifies pseudo-labeling indicators accordingly in the subsequent iteration. For advantageous classes, FedGGp employs adaptive thresholds to generate high-quality pseudo-labels, while for discriminated classes, it expands the number of pseudo-labels to ensure balanced model training. Moreover, soft consistency regularization is applied to broaden the boundary of pseudo-labels for some underrepresented classes, which are typically ambiguous during classifications. The experimental results on four different datasets demonstrate that FedGGp outperforms various state-of-the-art methods.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"294 ","pages":"Article 128667"},"PeriodicalIF":7.5,"publicationDate":"2025-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144489797","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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":"Two-dimensional logistic regression model of the white effect in printed media","authors":"Ivan Budimir,&nbsp;Igor Zjakić","doi":"10.1016/j.eswa.2025.128784","DOIUrl":"10.1016/j.eswa.2025.128784","url":null,"abstract":"<div><div>The White effect is one of the most pronounced psychophysical effects causing false color perception of observed elements. This effect frequently appears in graphic reproductions during the printing process. Thus, effectively managing the White effect is crucial for achieving desired visual outcomes and avoiding unintended ones. This study examines the behavior of the White psychophysical visual effect of induction and assimilation on various White grid variations in printed media. An experiment was conducted with 38 participants, and it was found that the effect is influenced by two parameters: grid coverage percentage and the brightness of rectangular elements. Focus is given to situations where the White effect is visible (Δ<em>L</em> &gt; 1), and a logistic regression model of visible differences is presented. This model allows for determining the probability of the White effect’s appearance in printed media, depending on grid coverage percentage and brightness parameters. The logistic regression model enables the identification and classification of White grid types in which the effect is noticeable. The study’s findings facilitate the identification of the White effect in printed media.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"293 ","pages":"Article 128784"},"PeriodicalIF":7.5,"publicationDate":"2025-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144491382","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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":"MSHF: Multi-sensor hierarchical fusion for UGV localization in unstructured environment","authors":"Jinwen Hu ,&nbsp;Jianyu Chen ,&nbsp;Mingwei Lv ,&nbsp;Zhao Xu ,&nbsp;Zhiwei Chen ,&nbsp;Junwei Han","doi":"10.1016/j.eswa.2025.128732","DOIUrl":"10.1016/j.eswa.2025.128732","url":null,"abstract":"<div><div>Multi-sensor fusion has been proven to be an efficient option for precise localization for unmanned ground vehicle (UGV) in GNSS-denied situations. This work presents a multi-sensor hierarchical fusion (MSHF) method for simultaneous localization and mapping (SLAM). Different from the existing methods, this work incorporates Odometer, Gear, and Steering wheel angle (OGS) information of a ground vehicle. In the first-level sensor fusion, the OGS data are fused with Inertial Measurement Unit (IMU) data to obtain a prior estimate of the vehicle state, where the adaptive extended Kalman filter is utilized to address the nonstationary measurement noise in the OGS data. The estimate is then fused with the one obtained by a conventional LiDAR based SLAM method in the second-level sensor fusion to provide a global optimal estimate, where the covariance intersection (CI) method is utilized for fusion of estimates with unknown correlation. The efficacy of the proposed method is demonstrated via a series of experiments and compared to the conventional algorithms using our self-generated datasets.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"294 ","pages":"Article 128732"},"PeriodicalIF":7.5,"publicationDate":"2025-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144518126","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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":"Protected template classification for iris biometrics","authors":"Qianrong Zheng ,&nbsp;Jianwen Xiang ,&nbsp;Rui Hao ,&nbsp;Songsong Liao ,&nbsp;Ling Dong ,&nbsp;Dongdong Zhao","doi":"10.1016/j.eswa.2025.128773","DOIUrl":"10.1016/j.eswa.2025.128773","url":null,"abstract":"<div><div>Although the wide application of biometrics has brought much convenience to the daily lives of people, it has also resulted in several security risks. Currently, many classical attack methods assume that the attacker has mastered all details of the template protection scheme. However, in practical environments, attackers often find it difficult to obtain the complete system information. Therefore, this paper proposes a two-stage classification model that can effectively classify template protection schemes. The model adopts a two-stage classification strategy: the first stage focuses on the overall classification of template protection schemes, whereas the second stage analyses the specific implementation details of each scheme and its related parameter settings in detail. Through this design, template protection schemes can still be effectively evaluated and attacked, even if the attacker does not have a full understanding of the specific details of the system. Further, this study explores the role of different classifiers in classification models in detail to help select the most appropriate classifier for improving the performance of the model. We compare several common classifiers and analyse their performance on different datasets. The experimental results show that the deep learning models (ResNet18 and DenseNet) outperform traditional machine learning models (LR, DT, RF, ADA) on all the datasets, with stable F1 scores greater than 0.90 on several testing sets, whereas the F1 scores of traditional machine learning models are generally low. In particular, on the CASIA-IrisV3-Interval and CASIA-IrisV4-Lamp datasets, DenseNet achieves F1 scores of 0.91-0.98, showing excellent generalisation ability. Further, the experiments show that deep learning models maintain high classification accuracies (F1 scores mostly higher than 0.90) even when the data sources of the template protection schemes differ, whereas the performance of some traditional models (e.g., ADA) fluctuates greatly on different datasets. This result suggests that with sophisticated guards, attackers can leverage the powerful classification capabilities of deep learning models to effectively analyse biometric data, thus providing a solid foundation for subsequent attack steps.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"294 ","pages":"Article 128773"},"PeriodicalIF":7.5,"publicationDate":"2025-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144513535","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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 novel image dataset for detecting and classifying mobility aid users","authors":"Sonia Dávila-Soberón,&nbsp;América Morales-Díaz,&nbsp;Mario Castelán","doi":"10.1016/j.eswa.2025.128697","DOIUrl":"10.1016/j.eswa.2025.128697","url":null,"abstract":"<div><div>When it comes to human identification as a computer vision task, artificial intelligence methods require extensive training to achieve good results. Large-scale image databases used for training and testing are easily available, however, disabled people still have poor representation in human datasets, making their visual identification hard to achieve. In this work, we introduce a new dataset based on wheelchair and cane users to fill the gap of in-the-wild images of disabled pedestrians and enable further research in the area. Additionally, we studied the effect of the dataset using transfer learning on state-of-the-art classification and detection models, training with combinations of the five classes available: wheelchair user, cane user, wheelchair, cane, and able-bodied person. Since this is the first work of its kind, we thoroughly analyzed the classification results across various image sizes and certainty thresholds. Furthermore, detection models trained with the new dataset were compared to those trained with a previously published mobility aid dataset through different evaluation metrics. Our results show high precision and certainty for both classification and detection, demonstrating the benefit the dataset has in the identification of mobility aid users and encouraging the inclusion of disabled people in the development of intelligent systems.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"293 ","pages":"Article 128697"},"PeriodicalIF":7.5,"publicationDate":"2025-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144472167","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}