International Journal of Intelligent Systems最新文献

{"title":"SEDNet: Real-Time Semantic Segmentation Algorithm Based on STDC","authors":"Sugang Ma,&nbsp;Ziyi Zhao,&nbsp;Zhiqiang Hou,&nbsp;Wangsheng Yu,&nbsp;Xiaobao Yang,&nbsp;Xiangmo Zhao","doi":"10.1155/int/8243407","DOIUrl":"https://doi.org/10.1155/int/8243407","url":null,"abstract":"<div>\u0000 <p>Recently, deep convolutional neural networks (DCNN) have been widely used in semantic segmentation tasks and have achieved high segmentation accuracy. However, most algorithms based on DCNN have high computational complexity, making them unsuitable for real-time segmentation. To solve this problem, this paper proposes a real-time semantic segmentation algorithm based on the STDC network. The algorithm adopts an “encoder–decoder” embedded in a U-shaped architecture to realize real-time segmentation while maintaining high accuracy. Following the encoder, a mixed pooling attention module is designed to expand the receptive field, enhancing the network model’s learning ability in complex scenarios. Then, a feature fusion module is used for combining features from different stages, and channel attention based on atrous convolution is employed to expand the receptive field and avoid dimensionality reduction learning. Finally, a Tversky-based detail loss function is used to encode more spatial details. The proposed algorithm was extensively tested on the challenging Cityscapes and CamVid datasets, and the experimental results showed that the proposed algorithm obtained 76.4% and 72.8% of mIoU, respectively. Meanwhile, our algorithm achieves 105.2 FPS and 165.6 FPS inference speed with a single NVIDIA GTX 1080Ti GPU, meeting the real-time segmentation requirements. The proposed algorithm can conduct real-time segmentation while maintaining high accuracy, achieving a good balance between accuracy and speed.</p>\u0000 </div>","PeriodicalId":14089,"journal":{"name":"International Journal of Intelligent Systems","volume":"2025 1","pages":""},"PeriodicalIF":5.0,"publicationDate":"2025-04-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1155/int/8243407","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143793781","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":"R-YOLOv5s: Improved YOLOv5s for Object Detection in Low-Light Environments","authors":"Yimeng Xia,&nbsp;Yuanmei Wang,&nbsp;Hao Luo,&nbsp;Shengzhe Liu,&nbsp;Tao Li","doi":"10.1155/int/8834271","DOIUrl":"https://doi.org/10.1155/int/8834271","url":null,"abstract":"<div>\u0000 <p>In response to the challenge of low detection accuracy exhibited by mainstream object detection models in low-light environments, this paper proposes a novel detection model named R-YOLOv5s. The model incorporates several key enhancements to address this issue. First, the SCI image enhancement algorithm is designed to preserve more target features and details. Next, a newly lightweight RepVIT backbone network is built to extract more image features; the global attention mechanism (GAM) is introduced to generate multiscale features that are more readily discernible, thereby enhancing the efficiency of feature capture. To significantly enhance the efficiency and precision of prediction box regression, a specialized loss function called SIoU loss is constructed. Results from experiments conducted on the ExDark dataset indicate notable improvements over the baseline model, with precision (P) increasing by 10%, recall (R) by 11%, and the mean average precision (mAP 0.5) by 13%. The newly devised R-YOLOv5s Model achieves higher detection accuracy in low-light environments, showcasing its effectiveness in addressing the challenges posed by such conditions.</p>\u0000 </div>","PeriodicalId":14089,"journal":{"name":"International Journal of Intelligent Systems","volume":"2025 1","pages":""},"PeriodicalIF":5.0,"publicationDate":"2025-04-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1155/int/8834271","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143793782","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":"Toward Diagnosis of Diseases Using Emerging Technologies: A Comprehensive Survey of the State of the Art in Metaverse","authors":"Nasim Aslani,&nbsp;Ali Garavand","doi":"10.1155/int/8820744","DOIUrl":"https://doi.org/10.1155/int/8820744","url":null,"abstract":"<div>\u0000 <p><b>Introduction:</b> The Metaverse, a rapidly growing technology in healthcare, is proving to be a game-changer in early disease detection and diagnosis. This study aimed to identify the latest scientific achievements in Metaverse, such as its effects, associated technologies, and obstacles for diagnosing diseases.</p>\u0000 <p><b>Methods:</b> In this review study, the scientific databases, including PubMed and Web of Science, were searched using related keywords. Related studies about using Metaverse in disease diagnosis were included according to inclusion and exclusion criteria. Data extraction was done using the data extraction form. The findings were summarized and reported in tables and figures according to the study objectives.</p>\u0000 <p><b>Results:</b> From 1706 retrieved articles, 28 studies were included according to inclusion and exclusion criteria. Most studies were conducted in 2023 (13 out of 28). 13 groups of specialists used Metaverse to diagnose diseases; oncologists and neurologists used it more than others. The most important technological aspects of the Metaverse were six main categories, including computer vision, artificial intelligence, virtual reality, blockchain, digital twin, and cloud computing. The Metaverse’s main effects in diagnostic interventions were 22 subcategories in five categories, including improving diagnosis, facilitating interactions, improving education, a better future, and uncertainty. The Metaverse’s role in improving diagnosis was particularly significant. The challenges of the Metaverse in diagnosis were seven subcategories: challenges related to the conducted studies, financial limitations, technological issues, structural issues, legal and ethical issues, its acceptance, and challenges about the nature of the Metaverse.</p>\u0000 <p><b>Conclusion:</b> Given the pivotal role of accurate diagnosis in patients treatment plans, the Metaverse’s potential in complex and challenging diagnoses is significant. However, it is important to note that this potential can only be fully realized through further research on utilizing the Metaverse in healthcare, specifically in disease diagnosis. This call for additional research is not just a suggestion but a necessity for the future of healthcare.</p>\u0000 </div>","PeriodicalId":14089,"journal":{"name":"International Journal of Intelligent Systems","volume":"2025 1","pages":""},"PeriodicalIF":5.0,"publicationDate":"2025-04-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1155/int/8820744","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143770442","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":"Traffic Flow Prediction Framework That Can Appropriately Process the Noise, Volatility, and Nonlinearity in Traffic Flow Data","authors":"Yingping Tang,&nbsp;Qiang Shang,&nbsp;Longjiao Yin,&nbsp;Hu Zhang","doi":"10.1155/int/1789796","DOIUrl":"https://doi.org/10.1155/int/1789796","url":null,"abstract":"<div>\u0000 <p>Accurate traffic flow prediction is crucial for improving transportation efficiency. To improve the accuracy of traffic flow prediction, we developed a traffic flow prediction framework—namely, traffic flow multicomponent network—that appropriately processes the noise, volatility, and nonlinearity in traffic flow data. This framework comprises three components: a factor selection component, traffic flow decomposition component, and traffic flow prediction component. The factor selection component considers the dynamic effects of weather-related, environmental, and spatiotemporal factors on traffic flow; it then extracts and analyzes factors exhibiting strong correlations with traffic flow. The traffic flow decomposition component optimizes the parameters of variational mode decomposition on the basis of the envelope entropy by using the sparrow search algorithm; it then transforms traffic flow into multiple intrinsic mode functions to enable accurate traffic flow prediction. Finally, the traffic flow prediction component constructs dynamic feature matrices by using a bidirectional gated recurrent unit model to identify relationships within the data. Moreover, it uses an attention mechanism to assign different weights to different features on the basis of the importance of these features to traffic flow prediction, thereby enabling the efficient processing of a large volume of data. The performance of the proposed framework was examined in experiments conducted on large volumes of traffic flow data with different time granularities. The results indicated that the proposed framework achieved high prediction accuracy and stability for various time granularities, data samples, dataset sizes, and noise conditions. Moreover, it generally outperformed existing traffic flow prediction models under all experimental conditions.</p>\u0000 </div>","PeriodicalId":14089,"journal":{"name":"International Journal of Intelligent Systems","volume":"2025 1","pages":""},"PeriodicalIF":5.0,"publicationDate":"2025-04-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1155/int/1789796","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143769971","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":"X-SCSANet: Explainable Stack Convolutional Self-Attention Network for Brain Tumor Classification","authors":"Rahad Khan,&nbsp;Rafiqul Islam","doi":"10.1155/int/1444673","DOIUrl":"https://doi.org/10.1155/int/1444673","url":null,"abstract":"<div>\u0000 <p>Brain tumors are devastating and shorten the patient’s life. It has an impact on the physical, psychological, and financial well-being of both patients and family members. Early diagnosis and treatment can reduce patients’ chances of survival. Detecting and diagnosing brain cancers using MRI scans is time-consuming and requires expertise in that domain. Nowadays, instead of traditional approaches to brain tumor analysis, several deep learning models are used to assist professionals and mitigate time. This paper introduces a stack convolutional self-attention network that extracts important local and global features from a freely available MRI scan dataset. Since the medical domain is one of the most sensitive fields, end-users should put their trust in the deep learning model before automating tumor classification. Therefore, the Grad-CAM method has been updated to better explain the model’s output. Combining local and global features improves brain tumor classification performance, with the suggested model reaching an accuracy of 96.44% on the relevant dataset. The proposed model’s precision, specificity, sensitivity, and F1-score are reported as 96.5%, 98.83%, 96.44%, and 96.4%, respectively. Furthermore, the layers’ insights are examined to acquire a deeper knowledge of the decision-making process.</p>\u0000 </div>","PeriodicalId":14089,"journal":{"name":"International Journal of Intelligent Systems","volume":"2025 1","pages":""},"PeriodicalIF":5.0,"publicationDate":"2025-03-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1155/int/1444673","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143741668","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":"Locality Sensitive Hashing-Based Deepfake Image Recognition for Athletic Celebrities","authors":"Bo Xiang,&nbsp;Qin Xie,&nbsp;Shuangzhou Bi,&nbsp;Edris Khezri","doi":"10.1155/int/1313970","DOIUrl":"https://doi.org/10.1155/int/1313970","url":null,"abstract":"<div>\u0000 <p>The rapid advancement of deepfake technology poses significant challenges to athletic celebrities, where altered or falsified media can impact athletes’ reputations, fan engagement, and the integrity of match broadcasting. This paper proposes a novel framework for deepfake image recognition for athletic celebrities using locality sensitive hashing (LSH). LSH, an efficient technique for high-dimensional nearest neighbor searches, is employed to detect and differentiate deepfake images from authentic media. By extracting high-dimensional features from images and videos using convolutional neural networks (CNNs), LSH is applied to hash similar content into clusters for quick and accurate deepfake detection. The proposed method is tested on real-world dataset, showing promising results in terms of accuracy and computational efficiency. This research highlights the importance of integrating advanced hashing techniques like LSH in safeguarding the authenticity of digital content and provides insights into future directions for deepfake detection mechanisms.</p>\u0000 </div>","PeriodicalId":14089,"journal":{"name":"International Journal of Intelligent Systems","volume":"2025 1","pages":""},"PeriodicalIF":5.0,"publicationDate":"2025-03-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1155/int/1313970","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143698999","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":"A Question-Aware Few-Shot Text-to-SQL Neural Model for Industrial Databases","authors":"Ren Li,&nbsp;Yu Chen,&nbsp;Hongyi Zhang,&nbsp;Jianxi Yang,&nbsp;Qiao Xiao,&nbsp;Shixin Jiang","doi":"10.1155/int/8124797","DOIUrl":"https://doi.org/10.1155/int/8124797","url":null,"abstract":"<div>\u0000 <p>Intelligent question answering over industrial databases is a challenging task due to the multicolumn context and complex questions. The existing methods need to be improved in terms of SQL generation accuracy. In this paper, we propose a question-aware few-shot Text-to-SQL approach based on the SDCUP pretrained model. Specifically, an attention-based filtering approach is proposed to reduce the redundant information from multiple columns in the industrial database scenario. We further propose an operator semantics enhancement method to improve the ability of identifying complex conditions in queries. Experimental results on the industrial benchmarks in the fields of electric energy and structural inspection show that the proposed model outperforms the baseline models across all few-shot settings.</p>\u0000 </div>","PeriodicalId":14089,"journal":{"name":"International Journal of Intelligent Systems","volume":"2025 1","pages":""},"PeriodicalIF":5.0,"publicationDate":"2025-03-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1155/int/8124797","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143707535","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":"A Lightweight Dynamic Hierarchical Neural Network Model and Learning Paradigm","authors":"Liping Liao,&nbsp;Junlong Lin,&nbsp;Wenjing Zhang,&nbsp;Jun Cai","doi":"10.1155/int/6833629","DOIUrl":"https://doi.org/10.1155/int/6833629","url":null,"abstract":"<div>\u0000 <p>In image analysis scenarios such as the Internet of Things and the metaverse, the introduction of federated learning (FL) is an effective solution to safeguard user data security and meet low-latency requirements during the machine learning process. However, due to the constrained computational power and memory of devices, facilitating the local training of complex models becomes challenging, thereby posing a significant obstacle to the application of FL. Consequently, a lightweight dynamic hierarchical neural network model and its learning paradigm are proposed in this study. Specifically, a lightweight compression method is designed based on enlarged receptive fields and separable convolutions to reduce redundancy in convolutional layer feature maps. A dynamic model partitioning method is devised, grounded in the Q-Learning reinforcement learning algorithm, to enable collaborative model training across multiple devices and enhance the utilization efficiency of device computing and storage resources. Furthermore, a hierarchical federated partition learning (HFSL) paradigm based on complete weight sharing is introduced to facilitate the compatibility of partitioned models with FL. Experimental results show that our lightweight model outperforms existing models in terms of accuracy, lightweight degree, and efficiency on image analysis tasks. Moreover, the proposed HFSL paradigm achieves performance comparable to centralized training.</p>\u0000 </div>","PeriodicalId":14089,"journal":{"name":"International Journal of Intelligent Systems","volume":"2025 1","pages":""},"PeriodicalIF":5.0,"publicationDate":"2025-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1155/int/6833629","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143690199","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":"CSI Acquisition in Internet of Vehicle Network: Federated Edge Learning With Model Pruning and Vector Quantization","authors":"Yi Wang,&nbsp;Junlei Zhi,&nbsp;Linsheng Mei,&nbsp;Wei Huang","doi":"10.1155/int/5813659","DOIUrl":"https://doi.org/10.1155/int/5813659","url":null,"abstract":"<div>\u0000 <p>The conventional machine learning (ML)–based channel state information (CSI) acquisition has overlooked the potential privacy disclosure and estimation overhead problem caused by transmitting pilot datasets during the estimation stage. In this paper, we propose federated edge learning for CSI acquisition to protect the data privacy in the Internet of vehicle network with massive antenna array. To reduce the channel estimation overhead, the joint model pruning and vector quantization algorithm for network gradient parameters is presented to reduce the amount of exchange information between the centralized server and devices. This scheme allows for local fine-tuning to adapt the global model to the channel characteristics of each device. In addition, we also provide theoretical guarantees of convergence and quantization error bound in closed form, respectively. Simulation results demonstrate that the proposed FL-based CSI acquisition with model pruning and vector quantization scheme can efficiently improve the performance of channel estimation while reducing the communication overhead.</p>\u0000 </div>","PeriodicalId":14089,"journal":{"name":"International Journal of Intelligent Systems","volume":"2025 1","pages":""},"PeriodicalIF":5.0,"publicationDate":"2025-03-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1155/int/5813659","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143638991","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":"A Novel Correlation Coefficient for Spherical Fuzzy Sets and Its Application in Pattern Recognition, Medical Diagnosis, and Mega Project Selection","authors":"Mehboob Ali,&nbsp;Wajid Ali,&nbsp;Ishtiaq Hussain,&nbsp;Rasool Shah","doi":"10.1155/int/9164932","DOIUrl":"https://doi.org/10.1155/int/9164932","url":null,"abstract":"<div>\u0000 <p>The correlation coefficient (CC) is a statistical measure that is very useful to quantify the strength and direction of the relationship between two variables, processes, or sets. The primary objective of this paper is to propose a novel CC explicitly tailored for spherical fuzzy sets (SFSs), aiming to address the limitations and drawbacks associated with existing CCs. Our approach employs statistical concepts to quantify the correlation between variables and datasets within the context of SFSs. We formulate our proposed CC for SFSs by incorporating variance and covariance as fundamental components. This innovative approach not only accurately quantifies the degree of correlation between two SFSs but also characterizes the nature of their relationship, whether it is positive, neutral, or negative. As a result, our CC yields numerical values within the range of [−1, 1]. In contrast, existing methods focus solely on measuring the degree of association between two SFSs and are unable to differentiate the nature of the relationship, especially in cases of inverse correlation. We conduct a comparison to evaluate the efficiency of our proposed scheme in comparison to existing techniques, using numerical examples to showcase the dominance of our method. The comparative results indicate that our proposed approach effectively addresses the limitations of existing methods and produces more reliable and precise results. Furthermore, we applied our method to address three real-world challenges in pattern recognition, medical diagnosis, and mega-project selection, demonstrating its practicality, advantages, and usefulness.</p>\u0000 </div>","PeriodicalId":14089,"journal":{"name":"International Journal of Intelligent Systems","volume":"2025 1","pages":""},"PeriodicalIF":5.0,"publicationDate":"2025-03-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1155/int/9164932","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143646129","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}