The rapid spread of wearable technologies has motivated the collection of a variety of signals, such as pulse rate, electrocardiogram (ECG), electroencephalogram (EEG), and others. As those devices are used to do so many tasks and store a significant amount of personal data, the concern of how our data can be exposed starts to gain attention as the wearable devices can become an attack vector or a security breach. In this context, biometric also has expanded its use to meet new security requirements of authentication demanded by online applications, and it has been used in identification systems by a large number of people. Existing works on ECG for user authentication do not consider a population size close to a real application. Finding real data that has a big number of people ECG’s data is a challenge. This work investigates a set of steps that can improve the results when working with a higher number of target classes in a biometric identification scenario. These steps, such as increasing the number of examples, removing outliers, and including a few additional features, are proven to increase the performance in a large data set. We propose a data improvement model for ECG biometric identification (user identification based on electrocardiogram—DETECT), which improves the performance of the biometric system considering a greater number of subjects, which is closer to a security system in the real world. The DETECT model increases precision from 78% to 92% within 1500 subjects, and from 90% to 95% within 100 subjects. Moreover, good False Rejection Rate (i.e., 0.064003) and False Acceptance Rate (i.e., 0.000033) were demonstrated. We designed our proposed method over PhysioNet Computing in Cardiology 2018 database.
4 March 2024 in Journal of Internet Services and Applications
Integrating big data and deep learning across various applications has significantly enhanced intelligence and efficiency in our daily lives. However, it also requires extensive data sharing, raising significant communication and privacy concerns. In this context, Federated Learning (FL) emerges as a promising solution to enable collaborative model training while preserving the privacy and autonomy of participating clients. FL facilitates collaborative model training by enabling data to be trained locally on devices, eliminating the need for individual information sharing among clients. A client selection mechanism strategically chooses a subset of participating clients to contribute to the model training in each learning round. However, an efficient selection of clients to participate in the training process directly impacts model convergence/accuracy and the overall communication load on the network. In addition, FL faces challenges when dealing with non-Independent and Non-Identically Distributed (non-IID) data, where the diversity in data distribution often leads to reduced classification accuracy. Hence, designing an efficient client selection mechanism in a scenario with non-IID data is essential, but it is still an open issue. This article proposes a Model Efficiency through Selective Federated Learning Algorithm called MESFLA. The mechanism employs a Centered Kernel Alignment (CKA) algorithm to search for similar models based on data weight or similarity between models, i.e., grouping participants with comparable data distributions or learning objectives. Afterward, MESFLA selects the clients with more relevance in each group based on data weight and entropy. Our comprehensive evaluation across multiple datasets, including MNIST, CIFAR-10, and CIFAR-100, demonstrates MESFLA's superior performance over traditional FL algorithms. Our results show an accuracy improvement and a minor loss in each client aggregation of the new global model sent to clients with a difference of 3 rounds using the Data Weight in comparison with the other selection methods.