IEEE Transactions on Consumer Electronics最新文献

{"title":"Hierarchical Dual-Strategy Unlearning for Biomedical and Healthcare Intelligence Using Imperfect and Privacy-Sensitive Medical Data","authors":"Yi Zhang;Chao Zhang;Zijian Li;Tianxiang Xu;Kunyu Zhang;Zhan Gao;Meinuo Li;Xiaohan Zhang;Qichao Qi;Bing Chen","doi":"10.1109/TCE.2026.3670687","DOIUrl":"https://doi.org/10.1109/TCE.2026.3670687","url":null,"abstract":"Large language models (LLMs) exhibit exceptional performance across diverse domains, yet their propensity to memorise training data poses substantial privacy risks, particularly within sensitive healthcare contexts where medical data is often imperfect, insufficiently labelled, or contains privacy-sensitive patient information. Here, we present a hierarchical dual-strategy framework for selective knowledge unlearning in medical LLMs that enables precise removal of specialised knowledge whilst preserving fundamental medical competencies, specifically addressing challenges in biomedical and healthcare intelligence using imperfect data. Our approach synergistically integrates geometric-constrained gradient updates, which selectively modulate parameters encoding target knowledge while safeguarding essential capabilities, with concept-aware token-level interventions that systematically distinguish between preservation-critical and unlearning-targeted tokens through a unified four-level medical concept hierarchy. Through comprehensive evaluation on the MedMCQA dataset targeting surgical knowledge removal, and cross-domain validation on the MHQA dataset encompassing anxiety, depression, trauma, and obsessive-compulsive disorder domains, we demonstrate that our method achieves superior selective unlearning performance (82.7% forgetting rate, 88.5% knowledge preservation) compared to existing approaches. Notably, our framework maintains robust privacy guarantees while requiring modification of only 0.1% of model parameters, establishing a paradigm for privacy-preserving medical AI systems that addresses regulatory compliance, hospital auditability, and ethical imperatives in clinical research environments. This work contributes to advancing biomedical and healthcare intelligence by providing an effective solution for auditing and managing imperfect and privacy-sensitive medical data in real-world clinical applications.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"72 3","pages":"6681-6690"},"PeriodicalIF":9.9,"publicationDate":"2026-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148836838","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"3D Non-Degenerate Hyperchaos: Design, Analysis, and Application in Image Encryption","authors":"Yiting Lin;Yunlong Liao;Wenjun Zeng;Yunan Wei;Donglong Chen;Xiaochen Yuan;Yupeng Li;Ugur Erkan;Abdurrahim Toktas;Chongfu Zhang;Yong Zhang;Suo Gao","doi":"10.1109/TCE.2026.3672135","DOIUrl":"https://doi.org/10.1109/TCE.2026.3672135","url":null,"abstract":"The widespread adoption of consumer electronics (CE), particularly in smart home ecosystems, has led to the proliferation of intelligent devices such as smart cameras, video doorbells, and smart door locks. While these devices enhance convenience and automation, they generate massive amounts of visual data containing sensitive personal information, thereby increasing the risk of privacy leakage and challenging traditional encryption methods in consumer IoT networks. To address this issue, a cryptanalysis-driven image encryption scheme via a chaos-based 3D S-box and genetic code is proposed, specifically designed to safeguard visual privacy in consumer applications. First, a novel dynamical architecture, formally termed 3D Non-degenerate Hyperchaos (3D-NDHC), is constructed to serve as a robust entropy source. Explicitly defined as a globally bounded volume-expanding hyperchaotic system, 3D-NDHC distinguishes itself from traditional dissipative maps by exhibiting global robust hyperchaos and scalable complexity, where the Lyapunov exponents grow logarithmically with control parameters, ensuring high unpredictability. Next, a dynamically generated three-dimensional S-box driven by these chaotic sequences is introduced to perform efficient image permutation. Subsequently, a genetic code method enhanced by cryptanalysis is applied to confuse the permuted image matrix, specifically addressing the security vulnerabilities often inherent in lightweight encryption algorithms. Finally, under the control of pseudo-chaotic sequences, a lightweight wave diffusion process is employed to spread the pixel values and produce the final ciphertext image. In contrast to conventional algorithms that rely on standard encryption techniques, the proposed approach integrates a chaos-driven 3D S-box with a cryptanalysis-enhanced genetic code scheme. This combination introduces a high level of confusion and diffusion, significantly increasing the difficulty for attackers attempting to illegally decrypt sensitive images from smart devices. Experimental results and security analyses demonstrate that the scheme effectively resists various cryptographic attacks and outperforms several recently proposed image encryption algorithms, offering a robust solution for protecting digital image privacy in smart home environments.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"72 3","pages":"5483-5494"},"PeriodicalIF":9.9,"publicationDate":"2026-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148837016","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Context Aware Conformer for Single Image Super-Resolution: Merging Global Attention With Incremental Features","authors":"Naveen Kumar Tiwari;Shyam Singh Rajput","doi":"10.1109/TCE.2026.3685737","DOIUrl":"https://doi.org/10.1109/TCE.2026.3685737","url":null,"abstract":"Single Image Super-Resolution (SISR) plays a key role in next-generation consumer electronics like smartphones, smart TVs, and VR devices, where clear and detailed images are essential. Transformer-based methods effectively extract global features but often struggle with fine detail extraction and adapting to changing contexts. To address this, we propose a Context-Aware Conformer Network (CACN) for SISR. The CACN combines Convolutional Neural Networks (CNNs) and transformer blocks, called Adaptive Context Transformer Blocks (ACTBs), due to their context-aware adaptation capabilities. These ACTBs are cascaded to form lightweight or deep CACN variants. Each ACTB consists of Cascaded Context Aware Transformers (CCATs) with an Adaptive Channel Attention and Spatial Attention Fusion to enhance transformer capabilities. Depthwise separable convolutions combined with transformer insights enhance significant channels and features. The ACTB also includes a Context Fusion Block with a Redundant Information Processing Unit to filter redundant information and fuse local and global contexts. Extensive experiments on benchmark datasets with scale factors <inline-formula> <tex-math>$times 2$ </tex-math></inline-formula>, <inline-formula> <tex-math>$times 3$ </tex-math></inline-formula>, and <inline-formula> <tex-math>$times 4$ </tex-math></inline-formula> show that CACN achieves state-of-the-art results with improvements in PSNR, SSIM, and LPIPS while maintaining lower computational overhead. These results demonstrate CACN’s ability to reconstruct high-resolution images efficiently.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"72 3","pages":"5924-5935"},"PeriodicalIF":9.9,"publicationDate":"2026-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148837094","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":"Efficient Cross-Domain Multi-Factor Authentication","authors":"Yongjiao Li;Zhenlei Wang;Zhu Cao","doi":"10.1109/TCE.2026.3658538","DOIUrl":"https://doi.org/10.1109/TCE.2026.3658538","url":null,"abstract":"Secure cross-domain authentication is crucial for protecting information exchange across terminals in different security domains. However, existing cross-domain authentication models suffer from inflexibility, high storage and computational consumption. In response to these problems, a cross-domain multi-factor authentication protocol with improved efficiency (CDAuth) is proposed. CDAuth employs a universally applicable multi-factor authentication approach, integrating pseudo-identity, attribute concealment, and biometric hiding. The authentication information sets can be freely added or modified to accommodate various access scenarios. This boosts privacy, flexibility, and applicability compared to traditional methods. Furthermore, multi-edge servers as Certification Authorities (CAs) store only the union of terminal attributes and a verification token per user in their subdomains, instead of each user’s attribute and biometric sets as in traditional schemes, significantly reducing storage consumption. Additionally, replacing the bilinear pairing operations with the native group operations, while maintaining security, significantly reduces computational consumption. Our analysis shows that CDAuth can resist common security attacks and is more efficient than the previous schemes.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"72 3","pages":"5897-5910"},"PeriodicalIF":9.9,"publicationDate":"2026-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148837110","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":"Text-Only Virtual Currency LLM for Consumer Electronics: Social Sentiment Anomaly Detection and Trading Guidance","authors":"Xiaotong Jiang;Zhiyu Zhang;Jun Wu","doi":"10.1109/TCE.2026.3683776","DOIUrl":"https://doi.org/10.1109/TCE.2026.3683776","url":null,"abstract":"Consumer participation in the virtual currency market is increasingly mediated through personal electronics. Social media sentiment plays an important role in this market, where collective sentiment can trigger significant price volatility. Conventional financial models often struggle to capture these dynamics, exposing consumers to potential trading risks. To address this issue, we propose VirInsight. The framework integrates fine-grained sentiment classification, transaction anomaly detection, and trading recommendation. The key novelty lies in the design of an anomaly detection mechanism that is aware of sentiment dynamics and links social media sentiment to consumer trading guidance. For the Vir-Llama model, we develop a five-class sentiment classification scheme. We also construct a new dataset, VirSent, which contains over 12,000 annotated tweets related to major virtual currencies, including Bitcoin, Ethereum, and Binance Coin. Experimental results show that the proposed sentiment-aware approach improves the accuracy of anomaly detection and enables the generation of practical trading recommendations. This research contributes a robust method for modeling sentiment in the virtual currency market and offers a new perspective on consumer behavior. It ultimately paves the way for intelligent financial applications designed to empower and protect the everyday consumer.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"72 3","pages":"5682-5693"},"PeriodicalIF":9.9,"publicationDate":"2026-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148837114","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":"EUPA: Embodied Uncertainty-Aware Multimodal Affect Perception for Consumer Electronic Systems","authors":"Xianxun Zhu;Rui Wang;Lucia Cascone;Xiaosong E;Hui Chen","doi":"10.1109/TCE.2026.3679325","DOIUrl":"https://doi.org/10.1109/TCE.2026.3679325","url":null,"abstract":"Multimodal sentiment analysis has achieved strong performance in controlled settings by integrating visual, acoustic, and textual signals. However, real-world consumer electronic systems operate under dynamic conditions with noisy, asynchronous, and partially observable inputs, challenging conventional models that assume reliable multimodal data. In this paper, we reformulate multimodal sentiment analysis as an embodied affect perception problem and propose EUPA, an uncertainty-aware framework that models modality reliability and performs confidence-guided temporal updates for online inference. EUPA dynamically regulates unreliable inputs and maintains a belief-based representation to support robust perception under evolving conditions. To evaluate this setting, we introduce an embodied evaluation protocol on CMU-MOSI and CMU-MOSEI with simulated modality dropout and temporal misalignment. Experimental results demonstrate improved robustness and temporal stability over existing methods, highlighting the potential of EUPA for real-world affect-aware applications.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"72 3","pages":"6508-6515"},"PeriodicalIF":9.9,"publicationDate":"2026-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148837065","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":"RTF-Net: A Lightweight Dual-View Framework for Fine-Grained Encrypted Short-Video Fingerprinting","authors":"Yinhao Xiao;Xiaoya Xu;Haoran Xie;Fu Lee Wang","doi":"10.1109/TCE.2026.3689649","DOIUrl":"https://doi.org/10.1109/TCE.2026.3689649","url":null,"abstract":"Short video platforms like TikTok have dominated global internet traffic. Despite the deployment of robust encryption protocols (e.g., TLS 1.3 and QUIC) to protect user confidentiality, side-channel information leaking from encrypted traffic remains a significant privacy threat. However, accurately identifying specific video content is increasingly challenging due to complex Adaptive Bitrate (ABR) algorithms, which homogenize traffic patterns, especially in high-motion viral content. Existing methods relying on simple statistical features or single-view deep learning models often struggle to capture the fine-grained fingerprints of specific videos under these conditions. In this paper, we propose a novel Multi-View Fusion Network, namely Rocket-Transformer Fusion (RTF-Net), that integrates time-domain convolutional features with frequency-domain attention mechanisms. Our approach utilizes MiniRocket to extract deterministic sequential patterns from traffic bursts, while simultaneously employing Fast Fourier Transform (FFT) combined with TinyBERT to capture global periodic dependencies in the frequency domain. We evaluate our method on a self-collected dataset comprising 2,000 encrypted traffic traces generated from 20 most viral short dramas (each exceeding 100 million views) on TikTok. Experimental results demonstrate that our method achieves a state-of-the-art classification accuracy of 91.33% using only the first 30 seconds of viewing traffic, significantly outperforming other baseline machine learning algorithms, showing that side-channel leakage is indeed a threat to short video platforms. To facilitate reproducibility and future research, we have open-sourced our dataset and code online.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"72 3","pages":"6227-6237"},"PeriodicalIF":9.9,"publicationDate":"2026-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148837089","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":"High Input Impedance Chopper-Stabilized Analog Front End Based on Variable-Capacitor Calibration for Bio-Consumer Electronics","authors":"Feng Yan;Jingjing Liu;Weijie Ge;Kangkang Sun;Xinying Su;Junhao He","doi":"10.1109/TCE.2026.3681220","DOIUrl":"https://doi.org/10.1109/TCE.2026.3681220","url":null,"abstract":"Analog front-end (AFE) circuits for bio-consumer electronics require ultra-high input impedance to ensure robust bio-potential acquisition under interference. This paper presents a chopping-stabilized AFE featuring an automatic input-impedance enhancement technique based on a positive feedback loop (PFL) with a continuously tunable variable-capacitor. An oscillation-detection-based calibration loop dynamically adjusts the PFL to compensate front-end parasitic capacitance, driving the system from an unstable condition to an optimal stable boundary with maximized input impedance while avoiding over-compensation. To improve calibration robustness, a frequency ripple suppression scheme is introduced, enabling precise control of the variable-capacitor independent of input frequency variations. The proposed variable-capacitance technique provides femtofarad-level resolution without relying on discrete capacitor arrays, achieving wide tuning range and high linearity. Fabricated in a <inline-formula> <tex-math>$0.18~mu $ </tex-math></inline-formula>m CMOS process, the AFE achieves an input impedance of 46.3 G<inline-formula> <tex-math>$Omega $ </tex-math></inline-formula> at 50 Hz, representing an <inline-formula> <tex-math>$83times $ </tex-math></inline-formula> improvement over conventional approaches. The prototype occupies 0.71 mm², achieves a mid-band gain of 40 dB, an input-referred noise of <inline-formula> <tex-math>$0.93~mu $ </tex-math></inline-formula>V_rms, and a power consumption of <inline-formula> <tex-math>$3.48~mu $ </tex-math></inline-formula>W. Experimental EEG and ECG measurements using comfortable dry electrodes validate the effectiveness of the proposed AFE for bio-consumer electronic applications.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"72 3","pages":"5247-5257"},"PeriodicalIF":9.9,"publicationDate":"2026-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148837144","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":"Community Detection in Multilayer Networks via Semi-Supervised Matrix Factorization","authors":"Rui Han;Amin Rezaeipanah","doi":"10.1109/TCE.2026.3683161","DOIUrl":"https://doi.org/10.1109/TCE.2026.3683161","url":null,"abstract":"Community detection in multilayer networks has gained increasing attention due to its ability to capture complex and heterogeneous relationships across diverse social platforms. However, most existing methods ignore user attributes and interlayer heterogeneity, reducing community detection robustness. To address these challenges, this study proposes the Consensus-Guided Semi-Supervised Matrix Factorization (CG-SSMF) algorithm, a novel framework that integrates structural, attribute-based, and cross-layer information into a unified learning process. The method integrates nonnegative matrix factorization with a consensus matrix to jointly capture topology, content, and overlapping users, applies semi-supervised clustering per layer to preserve heterogeneity, and finally fuses results through ensemble clustering into a unified global community structure. Experiments on real multilayer social networks show that the proposed method outperforms existing baselines and effectively leverages overlapping users to improve cross-layer integration.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"72 3","pages":"5628-5635"},"PeriodicalIF":9.9,"publicationDate":"2026-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148837157","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":"Co-Design of Control, Energy, and Buffer With Delay Determinism for Consumer Electronics in Industry 5.0","authors":"Lei Feng;Yalin Cao;Zhixiang Yang;Ying Li;Qiyang Zhang;Fanqin Zhou;Lijie Zhou;Lei Wei","doi":"10.1109/TCE.2025.3649084","DOIUrl":"https://doi.org/10.1109/TCE.2025.3649084","url":null,"abstract":"The deployment of wireless control systems in consumer electronics (CEs) is one of the critical components in Industry 5.0 to enable real-time monitoring and control with efficiency, flexibility, and reliability. This paper presents a novel model for the co-design of control, energy, and buffer management with delay determinism for CEs in Industry 5.0. To tackle the non-convex problem of minimizing long-term average joint cost, we propose an efficient resource allocation mechanism based on Lyapunov optimization theory and alternating optimization, utilizing control state information feedback. Specifically, the long-term problem is equivalently converted into an instantaneous optimization problem, which is then decomposed into two iteratively solvable subproblems, namely the buffer allocation problem and the bandwidth allocation problem. The subproblems are solved by branch and bound method and the interior point method, respectively. Simulation results demonstrate that the proposed mechanism outperforms four baseline mechanisms, achieving a balance between delay determinism, control stability, and efficient resource utilization.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"72 3","pages":"6594-6605"},"PeriodicalIF":9.9,"publicationDate":"2026-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148837190","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}