Electrocardiogram (ECG) signals exhibit unique electrical heart activity patterns that vary across individuals, offering significant potential as a biometric modality for human identification. This study introduces a ...
Electrocardiogram (ECG) signals exhibit unique electrical heart activity patterns that vary across individuals, offering significant potential as a biometric modality for human identification. This study introduces a novel non-fiducial framework for ECG-based biometric identification, integrating unsupervised features learning and deep learning methodologies. The proposed approach employs a 1D-Local Difference Pattern (1D-LDP) to extract discriminative features from raw ECG signals, capturing unique characteristics of individual heartbeat dynamics. Subsequently, a hybrid Stacked Autoencoder?Deep Belief Network (SAE-DBN) architecture is implemented to refine and optimize feature representation while enhancing the performance of the classification task, and effectively addressing challenges related to ECG signal variability and noise. Comparative analysis with existing Local Binary Pattern (LBP) variants and machine learning classifiers (e.g., SVM, KNN) confirms the robustness of the proposed approach in handling signal non-stationarity and artifacts. Experimental validation on two public databases, MIT-BIH Normal Sinus Rhythm and ECG-ID, showcases the framework’s advantages over traditional methods, achieving identification accuracies of $$\varvec{96.00\%}$$ and $$\varvec{94.80\%}$$ , respectively, along with a low equal error rate (EER) of $$\varvec{3.05\%}$$ . The results underscore its suitability for real-time applications, computational efficiency, and robust performance for real world physiological conditions.
Magnetic nanofibers are of great interest for applications like data transport and storage as well as in basic research. Especially bent nanofibers, which can unambiguously be produced by electrospinning, show a broad...
详细信息
Machine learning is a key component of many applications, among which decentralized learning has attracted wide attention because of its cost-effectiveness and high efficiency. However, decentralized learning is vulne...
详细信息
ISBN:
(数字)9798350379228
ISBN:
(纸本)9798350390780
Machine learning is a key component of many applications, among which decentralized learning has attracted wide attention because of its cost-effectiveness and high efficiency. However, decentralized learning is vulnerable to various Byzantine attacks. Existing defense models face challenges in defending against complex attacks and are often constrained by strict network topologies. To address this issue, a parameter aggregation rule based on reputation evaluation (RPV) is proposed in this paper. This rule establishes a reputation model for neighboring nodes and continuously updates it based on their distance performance during each iteration. By applying the reputation model of neighboring nodes to the Beta distribution, their reliability is obtained. Based on the reliability of neighboring nodes, the set of reliable nodes for this iteration and select parameter values from them for aggregation are determined. Finally, we gradually eliminate the influence of Byzantine nodes. Our experimental results on the MNIST dataset demonstrate that the proposed algorithm is resilient to attacks from any number of Byzantine nodes and outperforms previous defense models in terms of network topology constraints, training accuracy, and computational costs.
Speech emotion recognition (SER) aims to identify the speaker's emotional states in specific utterances accurately. However, existing methods still face feature confusion when attempting to recognize certain emoti...
详细信息
Speech emotion recognition (SER) aims to identify the speaker's emotional states in specific utterances accurately. However, existing methods still face feature confusion when attempting to recognize certain emotions because traditional acoustic feature extraction methods fail to capture dynamic emotional changes, blurring emotional boundaries. Additionally, existing classification networks (CNs) are constrained by fixed learning strategies, hindering their ability to capture subtle emotional nuances and resulting in label confusion. To address these two issues, we introduce 3D multiresolution modulation filtered cochleogram (MMCG) features by computing the deltas and delta-deltas of MMCG features to enhance the dynamic emotional changes and produce distinct emotional boundaries. We then customize a conditional emotion feature diffusion (CEFD) module, which progressively diffuses features based on emotional context to retain emotional nuances effectively and reduce reliance on conditioned information. In addition, a confidence filtering module is used to filter diffused features based on confidence-based posterior probabilities to ensure enhanced feature discrimination. We design a flexible training strategy named the progressive interleaved learning strategy (PILS) to learn further complex emotional nuances, which consists of two alternating stages: fine-tuning the CN parameters and supervising the CEFD output. Testing on the IEMOCAP, CASIA, and EMODB corpora demonstrates significant performance improvements in SER.
Drawing inspiration from the dynamics of biological groups, flocking behavior has captivated interest due to its adaptive, self-organizing, and resilient characteristics. However, the presence of numerous agents in an...
详细信息
ISBN:
(数字)9798350316339
ISBN:
(纸本)9798350316346
Drawing inspiration from the dynamics of biological groups, flocking behavior has captivated interest due to its adaptive, self-organizing, and resilient characteristics. However, the presence of numerous agents in an open network environment raises concerns regarding the potential for information leakage during flocking movements. To address this issue, we present a decentralized output mapping function, executed independently by each individual agent, to ensure the maintenance of flocking behavior, including velocity alignment, cohesion and collision avoidance. Additionally, to enhance resource efficiency and convergence speed, an event-triggered mechanism is employed to mitigate the excessive consumption of bandwidth and computation resources in traditional flocking control algorithms. Moreover, the designed finite-time controller facilitates rapid convergence of flocking behavior. Finally, the efficacy of our proposed approach is corroborated through numerical simulation experiments.
This paper explores the size-invariance of evaluation metrics in Salient Object Detection (SOD), especially when multiple targets of diverse sizes co-exist in the same image. We observe that current metrics are size-s...
This paper explores the size-invariance of evaluation metrics in Salient Object Detection (SOD), especially when multiple targets of diverse sizes co-exist in the same image. We observe that current metrics are size-sensitive, where larger objects are focused, and smaller ones tend to be ignored. We argue that the evaluation should be size-invariant because bias based on size is unjustified without additional semantic information. In pursuit of this, we propose a generic approach that evaluates each salient object separately and then combines the results, effectively alleviating the imbalance. We further develop an optimization framework tailored to this goal, achieving considerable improvements in detecting objects of different sizes. Theoretically, we provide evidence supporting the validity of our new metrics and present the generalization analysis of SOD. Extensive experiments demonstrate the effectiveness of our method. The code is available at https://***/Ferry-Li/SI-SOD.
Driver drowsiness electroencephalography (EEG) signal monitoring can timely alert drivers of their drowsiness status, thereby reducing the probability of traffic accidents. Graph convolutional networks (GCNs) have sho...
详细信息
The 2050 carbon-neutral vision spawns a novel energy structure revolution,and the construction of the future energy structure is based on equipment *** material,as the core of electrical power equipment and electrifie...
详细信息
The 2050 carbon-neutral vision spawns a novel energy structure revolution,and the construction of the future energy structure is based on equipment *** material,as the core of electrical power equipment and electrified transportation asset,faces unprecedented challenges and *** goal of carbon neutral and the urgent need for innovation in electric power equipment and electrification assets are first *** engineering challenges constrained by the insulation system in future electric power equipment/devices and electrified transportation assets are *** materials,including intelligent insulating material,high thermal conductivity insulating material,high energy storage density insulating material,extreme environment resistant insulating material,and environmental-friendly insulating material,are cat-egorised with their scientific issues,opportunities and challenges under the goal of carbon neutrality being *** the context of carbon neutrality,not only improves the understanding of the insulation problems from a macro level,that is,electrical power equipment and electrified transportation asset,but also offers opportunities,remaining issues and challenges from the insulating material *** is hoped that this paper en-visions the challenges regarding design and reliability of insulations in electrical equipment and electric vehicles in the context of policies towards carbon neutrality *** authors also hope that this paper can be helpful in future development and research of novel insulating materials,which promote the realisation of the carbon-neutral vision.
DC microgrids have gained in popularity in recent years due to features such as greater reliability, high efficiency, and control simplicity. Constant power loads (CPLs), on the other hand, are a problem and an effect...
详细信息
DC microgrids have gained in popularity in recent years due to features such as greater reliability, high efficiency, and control simplicity. Constant power loads (CPLs), on the other hand, are a problem and an effective cause of instability in DC MGs. To address the aforementioned issue, this paper discusses the stabilization of a buck power converter that supplies CPL via a regulated boost power converter. A robust sliding mode control is proposed, which stabilizes the output voltage and ensures the CPL's required power. The system is investigated in the context of external perturbations such as fluctuating input voltage and sudden change in the CPL. The robustness and efficacy of the proposed method have been demonstrated to confirm the system's overall stability. Simulation and experimentation are employed to verify the provided controller. The results verify the suggested controller's superiority.
automatic ECG-based arrhythmia detection plays a crucial role in the early prevention and diagnosis of heart diseases. Manual diagnosis of arrhythmias using ECG data is not only time-consuming but also prone to errors...
详细信息
暂无评论