IEEE Access最新文献

{"title":"A Few-Shot Knowledge Graph Completion Model With Neighbor Filter and Affine Attention","authors":"Hongfang Gong;Yingjing Ding;Minyi Ma","doi":"10.1109/ACCESS.2025.3529528","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3529528","url":null,"abstract":"In recent times, extensive scholarly focus has been directed towards the knowledge graph completion (KGC) due to the large number of triples that perform well in training tasks. However, the relations of realistic knowledge graphs (KGs) usually have long-tailed distributions, posing a great challenge in inferring new triples of task relationships from a limited number of triples. To tackle this challenge, methodologies for few-shot knowledge graph completion (FKGC) have been devised. These approaches employ a limited set of reference triples to forecast novel triples for various relations. However, existing FKGC approaches suffer from the drawbacks of not fully utilizing the structural information in KGs and ignoring the fine-grained information of interactions between entity pairs. In this paper, a FKGC model with neighbor filter and affine attention (NFAA) is proposed. The NFAA model filters 2-hop neighbors into a neighborhood scope for an entity aggregator and constructs a relation generator utilizing the affine attention mechanism to efficiently infer new triples for the few-shot relation task. Evaluations are performed using two publicly available benchmark datasets: NELL-one and Wiki-one. Experimental results validate the superiority of the NFAA model relative to several state-of-the-art approaches.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"12308-12320"},"PeriodicalIF":3.4,"publicationDate":"2025-01-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10840224","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142993509","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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":"Vehicle-to-Vehicle Communication Channel Estimator Based on Gate Recurrent Unit","authors":"Jun-Han Wang;He He;Kosuke Tamura;Shun Kojima;Jaesang Cha;Chang-Jun Ahn","doi":"10.1109/ACCESS.2025.3529768","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3529768","url":null,"abstract":"With the development of autonomous vehicle operation, vehicle-to-vehicle (V2V) communication plays an increasingly important role. However, in high-speed mobile environments, the channel has fast time-varying, which significantly decreases the property of channel estimation. On the other hand, the frame structure of the IEEE 802.11p standard contains a few number of pilots and a large pilot interval, which is not sufficient to track the rapidly changing channel environment accurately. In recent years, deep learning has been widely used for channel estimation. However, these methods typically perform poorly in high-speed mobility scenarios or have excessively high computational complexity. To alleviate such issues, this study proposes a channel estimation method by combining the sparrow search algorithm (SSA) and gated recurrent unit (GRU). In addition, this paper adds the attention mechanism to GRU to improve the robustness of the model. The computer simulation results confirm that, compared to traditional schemes, the proposed estimator can achieve a lower bit error rate (BER) and normalized mean squared error (NMSE). At the same time, the computational complexity of the algorithm has been reduced to some extent, allowing the estimator to complete the channel estimation faster. This study provides a useful reference for optimizing neural networks and thus improving the performance of channel estimators.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"12332-12342"},"PeriodicalIF":3.4,"publicationDate":"2025-01-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10840182","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142993628","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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":"Hybrid Movie Recommendation System With User Partitioning and Log Likelihood Content Comparison","authors":"Yongmao Yang;Kampol Woradit;Kenneth Cosh","doi":"10.1109/ACCESS.2025.3529515","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3529515","url":null,"abstract":"fIn the domain of recommendation systems, matrix decomposition is an effective strategy for mitigating issues related to sparsity and low space utilization. The Alternating Least Squares (ALS) method, in particular, stands out for its ability to process data in parallel, thereby enhancing computational efficiency. However, when dealing with an original rating matrix, the ALS method may inadvertently sacrifice some information, leading to increased error rates. To address these challenges, this paper proposes a novel hybrid model that integrates matrix factorization with additional features. Furthermore, it leverages weighted similarity measures and employs advanced log-likelihood text mining techniques. These innovations are designed to tackle cold-start problems and sparsity issues while compensating for information loss to mitigate errors. Under the premise that our model employs consistent evaluation metrics and datasets, comparative analysis against existing models from related literature demonstrates superior performance. Specifically, our model achieves a lower Root Mean Square Error (RMSE) of 0.82 and 0.88, alongside a higher F1 score of 0.94 and 0.92 in two datasets. Our proposed hybrid approach effectively addresses sparsity and mitigates information loss in matrix factorization, as demonstrated by these results.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"11609-11622"},"PeriodicalIF":3.4,"publicationDate":"2025-01-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10840180","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142993855","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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 Two-Point Association Tracking System Incorporated With YOLOv11 for Real-Time Visual Tracking of Laparoscopic Surgical Instruments","authors":"Nyi Nyi Myo;Apiwat Boonkong;Kovit Khampitak;Daranee Hormdee","doi":"10.1109/ACCESS.2025.3529710","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3529710","url":null,"abstract":"The application of real-time visual tracking in laparoscopic surgery has gained significant attention in recent years, driven by the growing demand for precise and automated surgical assistance. Instrument tracking, in particular, is critical for enhancing the safety and efficacy of minimally invasive surgery, where direct visibility is often limited. Real-time tracking of surgical instruments allows for more accurate maneuvering, reduces the risk of accidental tissue damage, and enables the development of advanced computer-assisted surgical systems. In this context, advancements in deep learning, particularly through detection models and modern tracking algorithms, have opened new avenues for addressing the challenges posed by real-time laparoscopic instrument tracking. However, according to the preliminary results, the existing combination of the detection model and tracking algorithm could not often handle the remaining challenges, including fast-motion speed, occlusion, overlapping, and close proximity of surgical instruments. This paper proposes a novel two-point association approach for surgical instrument tracking using a combination of YOLOv11 for object detection and refined ByteTrack for tracking. The proposed system is evaluated on a comprehensive dataset of surgical videos. The experimental results demonstrate superior performance in terms of segmentation accuracy (via F1-score), tracking robustness (via MOTA and HOTA), and real-time processing speed (via FPS). In order to validate the effectiveness of this research, real-time surgical instrument tracking is performed with the streaming of laparoscopic gynecologic surgery on a donated soft-tissue cadaver. The results indicate that the proposed system significantly improves the segmentation and tracking of surgical instruments, offering a reliable tool for enhancing intraoperative navigation and reducing the risk of surgical errors. This work contributes to the advancement of intelligent surgical systems, providing a foundation for further integration of machine learning techniques in the operating room.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"12225-12238"},"PeriodicalIF":3.4,"publicationDate":"2025-01-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10840191","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142993796","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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":"Classification Based on the Support Vector Machine for Determining Operational Targets for Controlling Electricity Usage With Conventional Meters: A Case Study of Industrial and Business Tariff Customers From PT PLN (Persero) Indonesia","authors":"Galih Arisona;Alief Pascal Taruna;Dwi Irwanto;Arif Bijak Bestari;Wildan Juniawan","doi":"10.1109/ACCESS.2025.3529295","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3529295","url":null,"abstract":"Electricity theft remains a significant challenge for PT PLN (Persero), Indonesia’s primary electricity provider, serving over 89 million customers as of 2023. The study focuses on industrial and business tariff customers, using a dataset from 2019 to 2023, which includes monthly consumption data from PLN’s postpaid customers across thirty operational units with the highest Electricity Use Control (P2TL) levels, covering customers with a maximum power of 6,600 VA. This approach differs from previous studies that rely on open or smart meter data, as this study uses conventional meters for data collection. In the dataset used for this research, losses from confirmed electricity theft amounted to approximately IDR 19 billion. This research aims to improve the detection of electricity theft through a machine learning-based model utilizing the Support Vector Machine (SVM) classification technique. The goal is to enhance the P2TL mechanism by accurately identifying potential targets for field verification. Various SVM kernels were tested, including Radial Basis Function (RBF), Linear, Polynomial (Poly), and Sigmoid, alongside classifiers such as SVM, Logistic Regression, Decision Tree, and Naïve Bayes. Results show that the SVM model, particularly with the RBF kernel, achieves optimal performance, with balanced precision and recall, especially with 30 months of historical data. This optimized model contributes to improving PLN’s operational efficiency, offering more accurate identification of electricity theft cases, leading to substantial financial savings by reducing losses from unpaid consumption. The findings offer practical benefits for reducing electricity theft and improving PLN’s monitoring system, especially in industrial and business sectors.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"12388-12398"},"PeriodicalIF":3.4,"publicationDate":"2025-01-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10840176","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142993850","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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":"Tailored Channel Pruning: Achieve Targeted Model Complexity Through Adaptive Sparsity Regularization","authors":"Suwoong Lee;Yunho Jeon;Seungjae Lee;Junmo Kim","doi":"10.1109/ACCESS.2025.3529465","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3529465","url":null,"abstract":"In deep learning, the size and complexity of neural networks have been rapidly increased to achieve higher performance. However, this poses a challenge when utilized in resource-limited environments, such as mobile devices, particularly when trying to preserve the network’s performance. To address this problem, structured pruning has been widely studied as it effectively reduces the network with little impact on performance. To enhance a model’s performance with limited resources, it is crucial to 1) utilize all available resources and 2) maximize performance within these limitations. However, existing pruning methods often require iterations of training and pruning or many experiments to find hyperparameters that satisfy a given budget or forcibly truncate parameters with a given budget, resulting in performance loss. To solve this problem, we propose a novel channel pruning method called Tailored Channel Pruning. Given a target budget (e.g., FLOPs and parameters), our method outputs a tailored network that automatically takes the budget into account during training and satisfies the target budget. During the integrated training and pruning process, our method adaptively controls sparsity regularization and selects important weights that can help maximize the accuracy within the target budget. Through various experiments on the CIFAR-10 and ImageNet datasets, we demonstrate the effectiveness of the proposed method and achieve state-of-the-art accuracy after pruning.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"12113-12126"},"PeriodicalIF":3.4,"publicationDate":"2025-01-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10840184","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142993164","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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":"Improved Low-Complexity Demodulation: Integrating Minimum-Mean-Square-Error and Maximum-Likelihood Detection for Image-Sensor-Based Visible Light Communication","authors":"Yuki Ohira;Shintaro Arai;Kengo Fujii;Tomohiro Yendo","doi":"10.1109/ACCESS.2025.3529498","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3529498","url":null,"abstract":"This study proposes an improved low-complexity signal demodulation method for a visible light communication (VLC) system with an image sensor as a receiver and light-emitting diode (LED) transmitters. Signal demodulation by VLC systems is challenging when image sensors are used as receivers because of defocusing. The traditional maximum likelihood detection (MLD)-based demodulation, although effective, incurs high computational costs, restricting its use in real-time applications. To address these limitations, this study introduces an improved error signal candidate selection algorithm that focuses on residual signals and channel matrix characteristics to refine the signal estimation result. Through iterative refinement using low-complexity detectors such as those based on zero-forcing (ZF), and minimum-mean-square error (MMSE), our method substantially narrows the search space for the MLD, thus striking a balance between demodulation performance and computational efficiency. The proposed method was validated through laboratory experiments to demonstrate substantial performance gains over existing techniques. Thus, this method advances the field of VLC demodulation, particularly for systems with image sensors.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"12138-12147"},"PeriodicalIF":3.4,"publicationDate":"2025-01-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10840233","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142993310","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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":"Semantic Image Synthesis via Class-Adaptive Cross-Attention","authors":"Tomaso Fontanini;Claudio Ferrari;Giuseppe Lisanti;Massimo Bertozzi;Andrea Prati","doi":"10.1109/ACCESS.2025.3529216","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3529216","url":null,"abstract":"In semantic image synthesis the state of the art is dominated by methods that use customized variants of the SPatially-Adaptive DE-normalization (SPADE) layers, which allow for good visual generation quality and editing versatility. By design, such layers learn pixel-wise modulation parameters to de-normalize the generator activations based on the semantic class each pixel belongs to. Thus, they tend to overlook global image statistics, ultimately leading to unconvincing local style editing and causing global inconsistencies such as color or illumination distribution shifts. Also, SPADE layers require the semantic segmentation mask for mapping styles in the generator, preventing shape manipulations without manual intervention. In response, we designed a novel architecture where cross-attention layers are used in place of SPADE for learning shape-style correlations and so conditioning the image generation process. Our model inherits the versatility of SPADE, at the same time obtaining state-of-the-art generation quality improving FID score by 5.6%, 1.4% and 3.4% on CelebMask-HQ, Ade20k and DeepFashion datasets respectively, as well as improved global and local style transfer. Code and models available at <uri>https://github.com/TFonta/CA2SIS</uri>.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"10326-10339"},"PeriodicalIF":3.4,"publicationDate":"2025-01-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10841835","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142993108","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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":"Achieving Faster and Smarter Chest X-Ray Classification With Optimized CNNs","authors":"Hassen Louati;Ali Louati;Khalid Mansour;Elham Kariri","doi":"10.1109/ACCESS.2025.3529206","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3529206","url":null,"abstract":"X-ray imaging is essential in medical diagnostics, particularly for identifying anomalies like respiratory diseases. However, building accurate and efficient deep learning models for X-ray image classification remains challenging, requiring both optimized architectures and low computational complexity. In this paper, we present a three-stage framework to enhance X-ray image classification using Neural Architecture Search (NAS), Transfer Learning, and Model Compression via filter pruning, specifically targeting the ChestX-Ray14 dataset. First, NAS is employed to automatically discover the optimal convolutional neural network (CNN) architecture tailored to the ChestX-Ray14 dataset, reducing the need for extensive manual tuning. Subsequently, we leverage transfer learning by incorporating pre-trained models, which enhances the model’s generalizability and reduces dependency on large volumes of labeled X-ray data. Finally, model compression through filter pruning, driven by evolutionary algorithms, trims redundant parameters to improve computational efficiency while preserving model accuracy. Experimental results demonstrate that this approach not only boosts classification accuracy on the ChestX-Ray14 dataset but also significantly reduces model size, making it suitable for deployment in resource-constrained environments, such as mobile and edge devices. This framework provides a practical, scalable solution to improve both the accuracy and efficiency of medical image classification.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"10070-10082"},"PeriodicalIF":3.4,"publicationDate":"2025-01-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10839370","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142993635","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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":"Optimization of Truck Platoon Size in Freeway Diverging Areas Based on Comprehensive Performance Evaluation","authors":"Zhifa Yang;Zongyao Li;Zhuo Yu;Wencai Sun;Jingjing Tian","doi":"10.1109/ACCESS.2025.3528460","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3528460","url":null,"abstract":"Truck platoons can enhance traffic throughput, achieve better fuel economy, and yield environmental benefits. However, in freeway diverging areas, excessively long truck platoons can cause a blocking effect on small vehicles exiting the freeway, thereby causing congestion and impacting the traffic flow in the section. Therefore, to balance the benefits between truck platoons and small vehicles, a comprehensive evaluation model was established, considering traffic efficiency, safety, and fuel economy. The analytic hierarchy process (AHP) method was employed to determine the importance of each indicator, resulting in a composite score. Using the Simulation of Urban Mobility (SUMO) platform, this paper examines the effects of truck platoon size (ranging from 2 to 15 trucks) on traffic efficiency, safety, and fuel consumption under varying conditions. The analysis considers low, medium, and high small vehicle traffic volumes of 750, 1200, and 1650 pcu/h/lane, as well as off-ramp probabilities for small vehicles of 10%, 20%, 30%, and 40%. Simulation results indicate that traffic efficiency initially increases and then decreases as the truck platoon size increases. An increase in the number of truck platoon members leads to a decrease in section safety, particularly noticeable under medium and high flow conditions. In three flow scenarios, truck platoon size of more than 5 trucks can achieve higher fuel economy. Taking an off-ramp probability of 10% for small vehicles as an example, the optimal truck platoon size ranges from 3 to 8 vehicles under low flow conditions, 2 to 7 vehicles under medium flow conditions, and 2 to 6 vehicles under high flow conditions. Hence, traffic managers in freeway diverging areas can utilize the findings of this study to select suitable truck-platoon size, enabling them to implement preemptive adjustment strategies for achieving optimal comprehensive performance.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"10299-10310"},"PeriodicalIF":3.4,"publicationDate":"2025-01-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10838520","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142993803","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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}