Blockchain-Based Reversible Data Hiding for Securing Medical Images.

Medical images carry a lot of important information for making a medical diagnosis. Since the medical images need to be communicated frequently to allow timely and accurate diagnosis, it has become a target for malicious attacks. Hence, medical images are protected through encryption algorithms. Recently, reversible data hiding on the encrypted images (RDHEI) schemes are employed to embed private information into the medical images. This allows effective and secure communication, wherein the privately embedded information (e.g., medical records and personal information) is very useful to the medical diagnosis. However, existing RDHEI schemes still suffer from low embedding capacity, which limits their applicability. Besides, such solution still lacks a good mechanism to ensure its integrity and traceability. To resolve these issues, a novel approach based on image block-wise encryption and histogram shifting is proposed to provide more embedding capacity in the encrypted images. The embedding rate is over 0.8 bpp for typical medical images. On top of that, a blockchain-based system for RDHEI is proposed to resolve the traceability. The private information is stored on the blockchain together with the hash value of the original medical image. This allows traceability of all the medical images communicated over the proposed blockchain network.

Hierarchical gated recurrent neural network with adversarial and virtual adversarial training on text classification.

Document classification aims to assign one or more classes to a document for ease of management by understanding the content of a document. Hierarchical attention network (HAN) has been showed effective to classify documents that are ambiguous. HAN parses information-intense documents into slices (i.e., words and sentences) such that each slice can be learned separately and in parallel before assigning the classes. However, introducing hierarchical attention approach leads to the redundancy of training parameters which is prone to overfitting. To mitigate the concern of overfitting, we propose a variant of hierarchical attention network using adversarial and virtual adversarial perturbations in 1) word representation, 2) sentence representation and 3) both word and sentence representations. The proposed variant is tested on eight publicly available datasets. The results show that the proposed variant outperforms the hierarchical attention network with and without using random perturbation. More importantly, the proposed variant achieves state-of-the-art performance on multiple benchmark datasets. Visualizations and analysis are provided to show that perturbation can effectively alleviate the overfitting issue and improve the performance of hierarchical attention network.