IEEE Transactions on Consumer Electronics最新文献

{"title":"PAGM: Partially Aligned Global and Marginal Multi-View Contrastive Clustering for Facial Recognition in Consumer Electronics","authors":"Dayu Hu;Tagrid Abdullah N. Alshalali;Xu Xu;Por Lip Yee;Zaffar Ahmed Shaikh;Jing Yang;Thippa Reddy Gadekallu","doi":"10.1109/TCE.2025.3626525","DOIUrl":"https://doi.org/10.1109/TCE.2025.3626525","url":null,"abstract":"Facial recognition has become a cornerstone of consumer electronics, underpinning critical functions such as device security, user authentication, and privacy-sensitive applications. As these systems become increasingly embedded in daily consumer devices like smartphones, smart home assistants, and wearables, ensuring their forensic reliability and preserving user privacy have become paramount. Despite significant advancements, traditional supervised learning methods often require substantial labeled data, making unsupervised approaches more desirable for scalable and privacy-preserving deployments. However, existing unsupervised facial recognition methods commonly assume perfect alignment across views, a condition that is rarely achieved in practical consumer electronics environments. Moreover, conventional techniques that combine multi-view information through concatenation or weighted fusion can introduce noise and compromise system integrity. To overcome these challenges, we introduce Partially-Aligned Global and Marginal Multi-View Contrastive Clustering (PAGM), a novel method specifically designed to handle partially aligned multi-view facial data. PAGM employs both global and marginal contrastive losses to enhance discriminative feature learning while maintaining privacy and data integrity. The global learning module is inspired by the transformer architecture, employing the Query (Q), Key (K), and Value (V) matrices to construct an attention mechanism that more effectively fuses cross-modal information. Meanwhile, the margin learning module optimizes the network by treating aligned samples as positive pairs and introducing a margin term to prevent trivial solutions. Extensive experimental results validate that PAGM significantly outperforms existing methods, particularly in challenging, unaligned multi-view facial recognition scenarios in consumer electronic devices, thereby offering a promising solution for privacy-preserving biometric systems.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"72 1","pages":"2570-2578"},"PeriodicalIF":10.9,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147557613","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":"IoT Applications in Energy-Efficient Consumer Electronics for Smart Cities","authors":"Zhaoteng Zeng;GuiLi Luo","doi":"10.1109/TCE.2025.3633246","DOIUrl":"https://doi.org/10.1109/TCE.2025.3633246","url":null,"abstract":"This paper presents an innovative approach to enhancing energy efficiency in smart consumer electronics by integrating Internet of Things (IoT) technologies with solar energy systems within the context of smart cities. The proposed framework leverages IoT-enabled devices for real-time energy monitoring, adaptive appliance scheduling, and environmental sensing to optimize load management dynamically. A hybrid optimization algorithm combining Grey Wolf Optimization, Particle Swarm Optimization, and Differential Evolution (HGPDO) is developed to minimize energy costs, reduce peak-to-average demand ratios, and lower carbon emissions, while simultaneously maintaining user comfort through intelligent scheduling and environmental control. Simulation results demonstrate significant reductions in energy consumption—up to 35%—and carbon footprint mitigation—up to 36%—across multiple scenarios involving grid-only, grid plus solar, and grid plus solar with energy storage systems. The study validates the efficacy of the integrated IoT-solar framework in balancing sustainability objectives with user preferences, highlighting its potential for widespread adoption in solar-powered smart cities. This work contributes to the development of scalable, adaptive, and user-centric energy management solutions that align with global efforts toward sustainable urbanization and clean energy utilization.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"72 1","pages":"2594-2600"},"PeriodicalIF":10.9,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147557639","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":"Context-Preserving and Sparsity-Aware Temporal Graph Network for Unified Face Forgery Detection","authors":"Koyya Deepthi Krishna Yadav;Ilaiah Kavati;Abhihkeshav Santosh Kurella;Savvy Jain;Yashsvini Katare","doi":"10.1109/TCE.2025.3633697","DOIUrl":"https://doi.org/10.1109/TCE.2025.3633697","url":null,"abstract":"Deepfakes and face forgeries continue to evolve, posing significant threats to consumer privacy, reputation, and public security. Existing deep learning approaches often focus on spatial inconsistencies but fail to capture relational context across facial regions and long-term temporal anomalies such as unnatural blinking or lip synchronization errors. We propose EDRL, a lightweight model that effectively captures spatiotemporal relational dependencies for robust face forgery detection. The architecture incorporates a Spatio-Temporal Attention (STA) module built upon a lightweight MC<inline-formula> <tex-math>$3_{18}~3$ </tex-math></inline-formula>D convolutional backbone, enabling motion-aware feature extraction and region-specific attention mapping. A Sparsity-Aware Edge Dropping Relation Learner (EDRL) constructs adaptive facial graphs by pruning redundant and less informative edges. A Temporal Adaptive Aggregation Network (TAAN) then aggregates frame-level features, ensuring that temporally significant representations are preserved even after edge pruning. Extensive evaluations show that EDRL achieves 98.4% accuracy on CASIA-FASD and reduces HTER by 6.2% on Replay-Attack compared to state-of-the-art baselines, while maintaining competitive results on digital forgery datasets. By enhancing robustness to diverse manipulations while reducing computational overhead, EDRL contributes towards a secure, lightweight, and deployable framework suitable for real-world applications.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"72 1","pages":"2579-2593"},"PeriodicalIF":10.9,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147557550","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":"IEEE Consumer Technology Society Officers and Committee Chairs","authors":"","doi":"10.1109/TCE.2026.3668403","DOIUrl":"https://doi.org/10.1109/TCE.2026.3668403","url":null,"abstract":"","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"72 1","pages":"C4-C4"},"PeriodicalIF":10.9,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11456299","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147557626","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":"Optimizing Tourist Experience Through Energy Aware Scheduling Using a Hybrid RNN-DQN Model","authors":"Gegencaobudao;Sulonggaowa;A Rongna;Siqintu Qi","doi":"10.1109/TCE.2026.3654974","DOIUrl":"https://doi.org/10.1109/TCE.2026.3654974","url":null,"abstract":"Optimizing the trade-off between user satisfaction and device longevity remains a critical challenge in the deployment of smart tourism services on resource-constrained consumer electronics. This paper proposes a Hybrid RNN-DQN framework designed to achieve energy-aware scheduling through the integration of temporal behavior prediction and deep reinforcement learning. We formalize the problem as a Partially Observable Markov Decision Process (POMDP) where a Recurrent Neural Network (RNN) extracts latent tourist intent from sequential mobility traces, which is then utilized by a Deep Q-Network (DQN) to derive optimal sensing policies. Unlike traditional heuristic-based energy management, our model employs a multi-objective reward function that dynamically balances service quality with power conservation. To handle the spatial complexity of urban environments, we introduce a Spatial-Temporal Grid transformation that enables 2D Convolutional Neural Networks (CNN) to extract topological features from Point-of-Interest (POI) networks. Extensive experimental evaluations on real-world datasets (Gowalla and Brightkite) demonstrate that the proposed hybrid architecture achieves a 32% increase in battery life and an 18% improvement in recommendation precision compared to state-of-the-art baselines. Statistical validation through paired t-tests and comprehensive ablation studies further confirm the synergy between the recurrent and reinforcement learning components. Our findings provide a scalable and generalizable solution for pervasive computing environments, setting a new benchmark for energy-efficient context-awareness in the Internet of Things (IoT) era.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"72 1","pages":"2601-2608"},"PeriodicalIF":10.9,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147557709","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":"Adaptive Edge Intelligence Framework for Resource-Constrained IoT in Consumer Electronics","authors":"Abdul Rehman;Mahmood Ul Hassan;Sadaqat Ali;Khalid Mahmood;Niyaz Ahmad Wani;Muhammad Shahid Anwar","doi":"10.1109/TCE.2025.3648062","DOIUrl":"https://doi.org/10.1109/TCE.2025.3648062","url":null,"abstract":"Edge computing is increasingly being identified as an essential tool for enabling real-time processing of data on resource-constrained Internet of Things (IoT) devices, especially within consumer electronics. Current deep learning architectures are less adaptable to resource fluctuations, as they fail to dynamically adapt and remain less efficient on resource-constrained devices. This research introduces as it improves decision support on resource-constrained Internet of Things devices by adapting complexity levels for real-time resource profiling. The proposed architecture integrates collaborative computing and peer-to-peer networking, while resource management is achieved through the context-aware component, which utilizes behavior analytics and environmental observations. The experiment showed efficiency improvement by 28% on the proposed architecture compared to other architectures, improving inference accuracy by 94.3%, and improving resource management and latency by 10-15% and 30%, respectively.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"72 1","pages":"2624-2631"},"PeriodicalIF":10.9,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147557562","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":"Human–Robot Shared Control Strategy Based on Operator Intent Prediction With Bayesian Filtering","authors":"Shiliang Shao;Weimin Hu;Xianyu Shi;Guangjie Han;Ting Wang;Chunhe Song;Chuxi Fang","doi":"10.1109/TCE.2025.3633938","DOIUrl":"https://doi.org/10.1109/TCE.2025.3633938","url":null,"abstract":"Robots are increasingly integrated into daily life as consumer electronics. Leveraging the Internet of Things, humans can remotely control robots that function as edge devices to perform various tasks. With advances in artificial intelligence, robots have acquired a certain level of intelligence and autonomy, enabling them to predict human intentions and assist decision-making processes. This study aims to enhance human-robot interactions by proposing an operator intent prediction (OIP) method based on Bayesian filtering. The OIP method integrates data from multiple observation sources and dynamically adjusts the target transition matrix, enabling real-time updates of the prediction probabilities for each target. Consequently, robots can accurately identify the fluctuating intentions of the operator and respond promptly. To seamlessly integrate human intent with robotic intelligence, a shared control strategy for decision-making based on intention prediction is proposed. This strategy introduces a dynamic shared control weight function that adjusts human and robot contributions in real time according to changes in the predicted probabilities of target intentions. This adaptive approach ensures robust performance in complex and unstructured environments. The proposed method was rigorously evaluated using multiple performance metrics, demonstrating its effectiveness and advantages in enhancing human-robot interactions, improving decision-making efficiency, and enabling robots to provide more intelligent and responsive assistance. In addition, we conducted physical experiments in real-world scenarios to further validate the effectiveness of our method.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"72 1","pages":"2609-2623"},"PeriodicalIF":10.9,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147557841","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":"IEEE Consumer Technology Society Board of Governors","authors":"","doi":"10.1109/TCE.2026.3668402","DOIUrl":"https://doi.org/10.1109/TCE.2026.3668402","url":null,"abstract":"","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"72 1","pages":"C3-C3"},"PeriodicalIF":10.9,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11456300","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147557711","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":"2025 Index IEEE Transactions on Consumer Electronics","authors":"","doi":"10.1109/TCE.2025.3648500","DOIUrl":"https://doi.org/10.1109/TCE.2025.3648500","url":null,"abstract":"","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"71 4","pages":"12542-12815"},"PeriodicalIF":10.9,"publicationDate":"2026-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11339251","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145929379","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":"IEEE Consumer Technology Society Board of Governors","authors":"","doi":"10.1109/TCE.2025.3633469","DOIUrl":"https://doi.org/10.1109/TCE.2025.3633469","url":null,"abstract":"","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"71 4","pages":"C3-C3"},"PeriodicalIF":10.9,"publicationDate":"2025-12-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11306169","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145778163","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}