Recent advances in deep learning have shown many successful stories in smart healthcare applications with data-driven insight into improving clinical institutions’ quality of care. Excellent deep learning models are heavily data-driven. The more data trained, the more robust and more generalizable the performance of the deep learning model. However, pooling the medical data into centralized storage to train a robust deep learning model faces privacy, ownership, and strict regulation challenges. Federated learning resolves the previous challenges with a shared global deep learning model using a central aggregator server. At the same time, patient data remain with the local party, maintaining data anonymity and security. In this study, first, we provide a comprehensive, up-to-date review of research employing federated learning in healthcare applications. Second, we evaluate a set of recent challenges from a data-centric perspective in federated learning, such as data partitioning characteristics, data distributions, data protection mechanisms, and benchmark datasets. Finally, we point out several potential challenges and future research directions in healthcare applications.

Reversible data hiding for encrypted media not only preserves the privacy of the media content but also can convey additional information during message transmission. Some studies advocate separability, that is, the message can be correctly extracted irrespective of whether the encrypted media has been decrypted. However, current research has focused on encrypted images. Urgent research is required on encrypted three-dimensional (3D) models. This paper proposes a separable reversible data hiding method based on spatial subdivision and space encoding for encrypted 3D models. A bounding volume is first constructed using the vertices with boundary values in the processing model. Each vertex coordinate value is then converted into a ratio (between 0 and 1) of the distance between the vertex and minimum boundary point to the side length of the bounding volume. The owner of the 3D model then uses a secret key to encrypt all ratios except those of the boundary vertices to obtain an encrypted 3D model of the same size as the original model. The spatial subdivision technique and a subdivision threshold are subsequently used to divide the bounding volume into a series of blocks and simultaneously control the vertex distortion. The secret message is embedded in the encrypted vertex by using the space encoding method with an embedding threshold. Experimental results indicate that the proposed algorithm enables high privacy, performs separable reversible data hiding, and has low computational complexity, high embedding capacity, and controllable distortion.

Reversible data hiding in encrypted media involves using algorithms to perform data encryption to improve the privacy of the original media and hide data for covert communication or access control. This study explores the feasibility of applying coordinate transformation and random sampling to enhance the embedding rate and total embedding capacity of separable reversible data hiding in encrypted polygonal models based on multiple most significant bit (multi-MSB) prediction and Huffman coding. We first transform each vertex coordinate value of the input model into a decimal value between 0 and 1 and then convert the value into numerous binary digits with a user-defined compression threshold. Thereafter, the random sampling concept is used to generate embeddable vertices with only partial neighboring vertices as a reference. Thus, the embedding rate and total embedding capacity can both be increased considerably. The multi-MSB prediction technique is adopted to obtain the embedding capacity of each embeddable vertex, each vertex coordinate value in the proposed study can have a respective embedding length. Finally, the auxiliary information compressed through Huffman coding and an encrypted secret message are embedded in the multi-MSB of each embeddable vertex coordinate value through bit substitution. The experimental results of this study indicate the feasibility of the proposed algorithm.

Reversible data hiding in encrypted images can simultaneously enhance the image privacy and preserve the message security for the purpose of covert communication and cloud data management. The algorithm extends the multi-MSB prediction and Huffman coding to propose the first reversible data hiding in encrypted HDR images in the Radiance RGBE format. Because of the high similarity among the exponent channel values of neighboring pixels, we directly applied multi-MSB prediction in our proposed preprocessing procedure, yielding considerably increased embedding capacity. We also proposed a novel image encryption method to maintain the characteristics of the images. Subsequently, a distortion-free data hiding algorithm, namely the homogeneity index modification algorithm, was added to further increase embedding capacity. The experimental results demonstrate the feasibility of the proposed algorithm and its ability to increase embedding capacity, embedding rate, and image privacy, support two data hiders, and enable reversibility and separability.