Pub Date : 2025-08-18DOI: 10.1109/TCE.2025.3584411
{"title":"IEEE Consumer Technology Society Board of Governors","authors":"","doi":"10.1109/TCE.2025.3584411","DOIUrl":"https://doi.org/10.1109/TCE.2025.3584411","url":null,"abstract":"","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"71 2","pages":"C3-C3"},"PeriodicalIF":10.9,"publicationDate":"2025-08-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11129002","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144867746","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}
Pub Date : 2025-08-11DOI: 10.1109/TCE.2025.3597357
Zheng Gao;Shancang Li;Muddear Iqbal
The integration of Generative AI (GAI) into consumer electronics (e.g., smart homes, wearables) introduces critical privacy risks as sensitive user data fuels personalized services. This paper proposes a homomorphic encryption-federated learning (HE-FL) framework that ensures end-to-end data confidentiality and decentralized model training. By combining HE’s encrypted computation with FL’s distributed architecture, the framework mitigates vulnerabilities in centralized systems while resisting probabilistic polynomial-time adversaries under the Learning With Errors (LWE) assumption. Evaluations on MNIST demonstrate a 3% accuracy trade-off (95.5% vs. 98.5% baseline) for robust privacy, reducing gradient inversion success to $leq 5%$ . Case studies in healthcare wearables and smart grids validate QoS-aware risk mitigation. Challenges in scalability and quantum-era security are addressed through edge-assisted optimizations and hybrid architectures, aligning with GDPR/CCPA compliance to foster trust in GAI-driven ecosystems.
将生成式人工智能（GAI）集成到消费电子产品（例如智能家居、可穿戴设备）中，由于敏感的用户数据为个性化服务提供了燃料，因此引入了关键的隐私风险。提出了一种保证端到端数据保密性和分散模型训练的同态加密联邦学习（HE-FL）框架。通过将HE的加密计算与FL的分布式架构相结合，该框架减轻了集中式系统中的漏洞，同时在有错误学习（LWE）假设下抵抗概率多项式时间对手。对MNIST的评估显示为3% accuracy trade-off (95.5% vs. 98.5% baseline) for robust privacy, reducing gradient inversion success to $leq 5%$ . Case studies in healthcare wearables and smart grids validate QoS-aware risk mitigation. Challenges in scalability and quantum-era security are addressed through edge-assisted optimizations and hybrid architectures, aligning with GDPR/CCPA compliance to foster trust in GAI-driven ecosystems.
{"title":"Enhancing Privacy in Generative AI-Enabled Consumer Electronics Using Homomorphic Encryption and Federated Learning","authors":"Zheng Gao;Shancang Li;Muddear Iqbal","doi":"10.1109/TCE.2025.3597357","DOIUrl":"https://doi.org/10.1109/TCE.2025.3597357","url":null,"abstract":"The integration of Generative AI (GAI) into consumer electronics (e.g., smart homes, wearables) introduces critical privacy risks as sensitive user data fuels personalized services. This paper proposes a homomorphic encryption-federated learning (HE-FL) framework that ensures end-to-end data confidentiality and decentralized model training. By combining HE’s encrypted computation with FL’s distributed architecture, the framework mitigates vulnerabilities in centralized systems while resisting probabilistic polynomial-time adversaries under the Learning With Errors (LWE) assumption. Evaluations on MNIST demonstrate a 3% accuracy trade-off (95.5% vs. 98.5% baseline) for robust privacy, reducing gradient inversion success to <inline-formula> <tex-math>$leq 5%$ </tex-math></inline-formula>. Case studies in healthcare wearables and smart grids validate QoS-aware risk mitigation. Challenges in scalability and quantum-era security are addressed through edge-assisted optimizations and hybrid architectures, aligning with GDPR/CCPA compliance to foster trust in GAI-driven ecosystems.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"71 3","pages":"9125-9131"},"PeriodicalIF":10.9,"publicationDate":"2025-08-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145510162","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}
The application of grasp pose detection technology is critical for improving robotic sorting efficiency in diverse electronic consumer goods, yet existing RGB-D based algorithms primarily optimize overall accuracy while neglecting significant performance imbalances across objects of varying scales, particularly for small-scale items. To address this limitation, we propose introduces a multi-scale consumer goods 6-DoF grasp pose balanced detection algorithm to overcome the multi-scale detection challenges. First, we develop a deep residual PointNet++ encoder that integrates depth information with hierarchical residual connections, enhancing global semantic feature extraction and mitigating information loss of small-scale consumer goods. Second, we design a multi-scale grasp width grouping feature extraction module that simultaneously captures multi-scale local geometric features under varying receptive fields through parallel dilated convolution branches, effectively resolving feature representation disparities across scales. Third, we introduce a multi-scale weight balanced loss function that dynamically adjusts optimization weights based on grasp width of consumer goods, addressing inherent data imbalance during training. Evaluated on the GraspNet-1Billion benchmark, our framework achieves remarkable improvements over GraspNet baseline, with accuracy gains of 18.71%, 9.35%, and 7.04% on small-scale objects across three test sets. Real-world experiments on a Franka Emika 7-DoF robot demonstrate a 91.79% success rate in cluttered multi-object scenarios. The robot grasping project and code are available at: https://upc-ghy.github.io/Franka-Grasp, and https://github.com/upc-ghy/GraspBalance.
{"title":"Multi-Scale Balanced Grasp Pose Detection for Robotic Sorting of Diverse Consumer Goods","authors":"Haiyuan Gui;Shanchen Pang;Xiao He;Xue Zhai;Nuanlai Wang;Sibo Qiao;Wenjing Yin;Sarra Ayouni;Mohamed Maddeh","doi":"10.1109/TCE.2025.3596898","DOIUrl":"https://doi.org/10.1109/TCE.2025.3596898","url":null,"abstract":"The application of grasp pose detection technology is critical for improving robotic sorting efficiency in diverse electronic consumer goods, yet existing RGB-D based algorithms primarily optimize overall accuracy while neglecting significant performance imbalances across objects of varying scales, particularly for small-scale items. To address this limitation, we propose introduces a multi-scale consumer goods 6-DoF grasp pose balanced detection algorithm to overcome the multi-scale detection challenges. First, we develop a deep residual PointNet++ encoder that integrates depth information with hierarchical residual connections, enhancing global semantic feature extraction and mitigating information loss of small-scale consumer goods. Second, we design a multi-scale grasp width grouping feature extraction module that simultaneously captures multi-scale local geometric features under varying receptive fields through parallel dilated convolution branches, effectively resolving feature representation disparities across scales. Third, we introduce a multi-scale weight balanced loss function that dynamically adjusts optimization weights based on grasp width of consumer goods, addressing inherent data imbalance during training. Evaluated on the GraspNet-1Billion benchmark, our framework achieves remarkable improvements over GraspNet baseline, with accuracy gains of 18.71%, 9.35%, and 7.04% on small-scale objects across three test sets. Real-world experiments on a Franka Emika 7-DoF robot demonstrate a 91.79% success rate in cluttered multi-object scenarios. The robot grasping project and code are available at: <uri>https://upc-ghy.github.io/Franka-Grasp</uri>, and <uri>https://github.com/upc-ghy/GraspBalance</uri>.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"71 3","pages":"9157-9169"},"PeriodicalIF":10.9,"publicationDate":"2025-08-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145486517","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}
Pub Date : 2025-08-01DOI: 10.1109/TCE.2025.3594723
Jing Zhang;Ya-Juan Yang
Consumer Electronics Data Fusion (CEDF) integrates diverse and heterogeneous consumer-generated data, crucial for advancing innovation, user experience, and decision-making in Consumer Electronics (CE). However, the challenges of multi-modal data diversity and sparse data structures complicate the fusion process. This article proposes a novel Multi-modal Multi-objective Evolutionary Algorithm (MMOEA) that emphasizes the construction and integration of Large Models (LMs) for adaptive CEDF. First, we design two metrics to balance alignment accuracy and conservativity, forming a robust multi-objective optimization model. Second, we seamlessly integrate LMs into the MMOEA, dynamically constructing task-specific LM ensembles that harness complementary model strengths to enhance semantic comprehension and scalability. Third, Auxiliary Matrices (AMs) are incorporated to guide the search process, ensuring convergence and diversity in solution spaces. Extensive experiments on the OAEI conference dataset and ten real-world CEDF task pairs validate the MMOEA’s effectiveness in producing high-quality, diverse solutions and offering decision-makers robust fusion strategies for complex CE systems.
{"title":"Adaptive Large Model Building for Consumer Electronics Data Fusion via Multi-Modal Multi-Objective Evolutionary Algorithm","authors":"Jing Zhang;Ya-Juan Yang","doi":"10.1109/TCE.2025.3594723","DOIUrl":"https://doi.org/10.1109/TCE.2025.3594723","url":null,"abstract":"Consumer Electronics Data Fusion (CEDF) integrates diverse and heterogeneous consumer-generated data, crucial for advancing innovation, user experience, and decision-making in Consumer Electronics (CE). However, the challenges of multi-modal data diversity and sparse data structures complicate the fusion process. This article proposes a novel Multi-modal Multi-objective Evolutionary Algorithm (MMOEA) that emphasizes the construction and integration of Large Models (LMs) for adaptive CEDF. First, we design two metrics to balance alignment accuracy and conservativity, forming a robust multi-objective optimization model. Second, we seamlessly integrate LMs into the MMOEA, dynamically constructing task-specific LM ensembles that harness complementary model strengths to enhance semantic comprehension and scalability. Third, Auxiliary Matrices (AMs) are incorporated to guide the search process, ensuring convergence and diversity in solution spaces. Extensive experiments on the OAEI conference dataset and ten real-world CEDF task pairs validate the MMOEA’s effectiveness in producing high-quality, diverse solutions and offering decision-makers robust fusion strategies for complex CE systems.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"71 4","pages":"12464-12478"},"PeriodicalIF":10.9,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145778257","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}
Training deep learning models requires large and diverse datasets. However, some fields, such as medical imaging and defect detection, encounter data collection and labeling issues, privacy concerns, and high annotation costs. This paper presents an unsupervised learning method combined with the attention mechanism to optimize the classification model for domain adaptation. Based on an autoencoder architecture, the attention mechanism ensures that teacher and student models focus on the same features within the same image during simultaneous training, thus improving classification accuracy. The proposed approach reduces overfitting by learning homogeneity and heterogeneity among categories, which enhances the model’s generalization on limited labeled data through few-shot learning. Experimental results show that our method improves segmentation accuracy by 20% mIoU over the baseline method on benchmark datasets, demonstrating its effectiveness in few-shot segmentation tasks.
{"title":"Improved Semantic Segmentation With Large-Scale Attention-Based Self-Supervised Few-Shot Learning","authors":"Jing-Ming Guo;Wei-Tse Wang;Yi-Chong Zeng;Zhen-Yu Chen","doi":"10.1109/TCE.2025.3587291","DOIUrl":"https://doi.org/10.1109/TCE.2025.3587291","url":null,"abstract":"Training deep learning models requires large and diverse datasets. However, some fields, such as medical imaging and defect detection, encounter data collection and labeling issues, privacy concerns, and high annotation costs. This paper presents an unsupervised learning method combined with the attention mechanism to optimize the classification model for domain adaptation. Based on an autoencoder architecture, the attention mechanism ensures that teacher and student models focus on the same features within the same image during simultaneous training, thus improving classification accuracy. The proposed approach reduces overfitting by learning homogeneity and heterogeneity among categories, which enhances the model’s generalization on limited labeled data through few-shot learning. Experimental results show that our method improves segmentation accuracy by 20% mIoU over the baseline method on benchmark datasets, demonstrating its effectiveness in few-shot segmentation tasks.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"71 3","pages":"9188-9192"},"PeriodicalIF":10.9,"publicationDate":"2025-07-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145486501","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}
According to the World Health Organization (WHO), COVID-19 has resulted in approximately 7 million deaths worldwide, posing a severe threat to public health. Accurately predicting COVID-19 infection trends can assist governments in developing strategies to mitigate the impact. This paper introduces a novel hybrid machine learning model, RLG-SIR-Net, proposed for predicting daily confirmed COVID-19 cases. DLinear is used to decompose time series data, obtaining a trend sequence and a residual sequence. The dynamic graph learning module can construct a dynamic graph from the trend sequence. Then, a graph convolutional network is adopted to extract correction information from the dynamic graph and the residual sequence. Finally, the correction information is employed to enhance the predictive performance of the SIR model. COVID-19 datasets containing data on four countries and five baseline models were used to validate the predictive performance of RLG-SIR-Net. Experimental results show that RLG-SIR-Net outperforms the other baseline models in long-term forecasting of COVID-19 infections.
{"title":"A Graph SIR Network Based on Dynamic Graph Structures and Residual Learning for Epidemic Prediction","authors":"Lingfeng Miao;Yufan Chen;Jiawei Wang;Choujun Zhan;Xuejiao Zhao","doi":"10.1109/TCE.2025.3587036","DOIUrl":"https://doi.org/10.1109/TCE.2025.3587036","url":null,"abstract":"According to the World Health Organization (WHO), COVID-19 has resulted in approximately 7 million deaths worldwide, posing a severe threat to public health. Accurately predicting COVID-19 infection trends can assist governments in developing strategies to mitigate the impact. This paper introduces a novel hybrid machine learning model, RLG-SIR-Net, proposed for predicting daily confirmed COVID-19 cases. DLinear is used to decompose time series data, obtaining a trend sequence and a residual sequence. The dynamic graph learning module can construct a dynamic graph from the trend sequence. Then, a graph convolutional network is adopted to extract correction information from the dynamic graph and the residual sequence. Finally, the correction information is employed to enhance the predictive performance of the SIR model. COVID-19 datasets containing data on four countries and five baseline models were used to validate the predictive performance of RLG-SIR-Net. Experimental results show that RLG-SIR-Net outperforms the other baseline models in long-term forecasting of COVID-19 infections.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"71 3","pages":"9185-9187"},"PeriodicalIF":10.9,"publicationDate":"2025-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145486504","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}
Pub Date : 2025-06-17DOI: 10.1109/TCE.2025.3561644
{"title":"IEEE Consumer Technology Society Board of Governors","authors":"","doi":"10.1109/TCE.2025.3561644","DOIUrl":"https://doi.org/10.1109/TCE.2025.3561644","url":null,"abstract":"","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"71 1","pages":"C3-C3"},"PeriodicalIF":4.3,"publicationDate":"2025-06-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11038973","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144308204","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}
Pub Date : 2025-06-17DOI: 10.1109/TCE.2025.3561646
{"title":"IEEE Consumer Technology Society Officers and Committee Chairs","authors":"","doi":"10.1109/TCE.2025.3561646","DOIUrl":"https://doi.org/10.1109/TCE.2025.3561646","url":null,"abstract":"","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"71 1","pages":"C4-C4"},"PeriodicalIF":4.3,"publicationDate":"2025-06-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11038963","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144308480","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}
The rapid advancements in AI-Generated Content (AIGC) have revolutionized various domains, including consumer electronics and healthcare technologies. AIGC’s ability to generate high-quality text, images, and videos within seconds has reshaped human-computer interactions, from intelligent customer service to immersive virtual experiences. More importantly, its application in Healthcare Electronics (HE) has opened new frontiers, facilitating automated diagnostics, medical data synthesis, and intelligent healthcare predictions.
{"title":"Guest Editorial AI-Generated Content Empowered Healthcare Electronics","authors":"Gwanggil Jeon;Joel Rodrigues;Shiping Wen;Junxin Chen;Nan Ji;Abdellah Chehri","doi":"10.1109/TCE.2025.3553026","DOIUrl":"https://doi.org/10.1109/TCE.2025.3553026","url":null,"abstract":"The rapid advancements in AI-Generated Content (AIGC) have revolutionized various domains, including consumer electronics and healthcare technologies. AIGC’s ability to generate high-quality text, images, and videos within seconds has reshaped human-computer interactions, from intelligent customer service to immersive virtual experiences. More importantly, its application in Healthcare Electronics (HE) has opened new frontiers, facilitating automated diagnostics, medical data synthesis, and intelligent healthcare predictions.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"71 1","pages":"1319-1321"},"PeriodicalIF":4.3,"publicationDate":"2025-06-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11038991","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144323198","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}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：