Multimedia significantly enhances modern healthcare by facilitating the analysis and sharing of diverse data, including medical images, videos, and sensor data. Integrating AI for multimedia data classification shows promise in improving healthcare services, data analysis, and decision-making. However, ensuring privacy in AI-integrated healthcare systems remains a challenge, especially with data continuously transmitted over networks. Synchronous Federated Learning (FL) is designed to address these privacy concerns by allowing end devices to collaboratively train a machine learning model without sharing data. Nonetheless, FL alone does not fully resolve privacy issues and faces efficiency challenges, particularly with devices of varying computational capabilities. In this article, we introduce an Asynchronous Partial Privacy-preserving Split- Federated Learning (APP-SplitFed) approach for smart healthcare systems. This method reduces computational demands on resource-limited devices and uses a weight-based aggregation method to allow devices of differing computational power to contribute effectively, ensuring optimal model performance and rapid convergence. Additionally, we incorporate a secure aggregation method to prevent adversaries from identifying individual models owned by healthcare institutions.

The consumer health industry has witnessed a transformative revolution driven by the Internet of Medical Things (IoMT) and the Metaverse. Additionally, Artificial Intelligence (AI) further enhances this evolution by unlocking the potential for data-driven insights. Yet, there is an imperative requirement to safeguard consumer healthcare data, particularly with the challenges posed by resource-constrained consumer healthcare devices and the complex, non-Independent and Identically Distributed (non-IID) nature of data flows between the Metaverse and healthcare entities. The introduction of Digital Twin (DTs) has emerged as a groundbreaking innovation in the consumer health industry, offering real-time, dynamic digital representations of physical entities and providing invaluable insights and situational awareness. To address the multifaceted challenges, this paper introduces Decentralized SplitFed Learning (DSFL), a collaborative AI learning technique designed for deployment within the resource-constrained consumer health industry environment. DSFL optimizes resource utilization and mitigates computation burdens through decentralized model distribution. Furthermore, DSFL tackles the issue of non-IID data distribution within the dynamic Metaverse-driven consumer healthcare ecosystem by incorporating a clustering-based mechanism, which enhances learning performance by aligning the DTs' representations of healthcare consumers' real-time data statistics with the everchanging data streams within the Metaverse.

The advancement of internet and communication technologies has led to the era of Industry 4.0. This shift is followed by healthcare industries creating the term Healthcare 4.0. In Healthcare 4.0, the use of Internet of Things-enabled medical imaging devices for early disease detection has enabled medical practitioners to increase healthcare institutions' quality of service. However, Healthcare 4.0 is still lagging in artificial intelligence and big data compared to other Industry 4.0 due to data privacy concerns. In addition, institutions' diverse storage and computing capabilities restrict institutions from incorporating the same training model structure. This article presents a secure multiparty computation-based ensemble federated learning with blockchain that enables heterogeneous models to collaboratively learn from healthcare institutions' data without violating users' privacy. Blockchain properties also allow the party to enjoy data integrity without trust in a centralized server while also providing each healthcare institution with auditability and version control capability.

The advancement of Industrial Internet of Things (IIoT) technology has resulted in the fourth industrial revolution, or Industry 4.0, enabling industries to enhance productivity. However, despite the benefits, there remain significant challenges, such as resource heterogeneity, communication efficiency, and data privacy, that limit the applications of IIoT in privacy-sensitive domains like healthcare. In order to protect data privacy, Federated Learning (FL) has been suggested as a solution, involving the sharing of model parameters rather than data itself. Current FL applications, however, still struggle with cost efficiency, especially when IIoT devices with heterogenous resources are involved. To address this, this paper proposes Digital Twin (DT) enabled Asynchronous SplitFed Learning (DT-ASFL) for classification tasks in the e-healthcare system over mobile networks. We first develop SplitFed Learning to introduce communication efficiency in the e-healthcare system, sending only extracted features during the learning process instead of the entire learning model. This enables resource-constrained devices to participate in the learning process by allowing the participants to train a partial learning model. DT is then employed to provide real-time statuses of IIoT devices deployed in the system, enabling asynchronous model updates in SplitFed Learning. The experimental results demonstrate the efficacy of DT-ASFL compared to the existing methods.

We propose a privacy-preserving ensemble infused enhanced deep neural network (DNN)-based learning framework in this article for Internet of Things (IoT), edge, and cloud convergence in the context of healthcare. In the convergence, the edge server is used for both storing IoT produced bioimage and hosting DNN algorithm for local model training. The cloud is used for ensembling local models. The DNN-based training process of a model with a local data set suffers from low accuracy, which can be improved by the aforementioned convergence and ensemble learning. The ensemble learning allows multiple participants to outsource their local model for producing a generalized final model with high accuracy. Nevertheless, ensemble learning elevates the risk of leaking sensitive private data from the final model. The proposed framework presents a differential privacy-based privacy-preserving DNN with transfer learning for a local model generation to ensure minimal loss and higher efficiency at the edge server. We conduct several experiments to evaluate the performance of our proposed framework.

The advancements in positioning technologies have led to the emergence of various location-based services, resulting in a drastic increase in location-based data generation, producing big-data. Location data are often linked with user privacy, as they can reveal sensitive information such as the places visited by a person. Moreover, most location-based services involve resource-constrained devices, needing lightweight data processing approaches. Due to these reasons, privacy and efficiency have been two of the primary components of location-based data processing. The existing approaches do not study both issues in the same setting. Consequently, current methods fail to provide efficient privacy preservation solutions towards location-based data stream processing. To address these issues, we investigate the effective integration of edge computing, cloud computing and differential privacy for location-based data clustering, which is an essential area in service recommendations (e.g. recommending the closest hotels to a particular location). In the proposed setup, we use local differential privacy to ensure user privacy. Next, we apply edge-based clustering on the differentially private input data using mobile edge devices. Next, the centroids of the clusters are collected at a cloud server to generate final clustering in a privacy-preserving manner. Our experiments show that the proposed approach provides maximum accuracy of 90% on lower privacy budgets (e.g. ɛ = 0.45-0.5).