DeepKeyGen: A Deep Learning-Based Stream Cipher Generator for Medical Image Encryption and Decryption

Yi Ding,Fuyuan Tan,Zhen Qin,Mingsheng Cao,Kim-Kwang Raymond Choo,Zhiguang Qin

DOI: https://doi.org/10.1109/tnnls.2021.3062754

IF: 14.255

2021-01-01

IEEE Transactions on Neural Networks and Learning Systems

Abstract:The need for medical image encryption is increasingly pronounced, for example, to safeguard the privacy of the patients' medical imaging data. In this article, a novel deep learning-based key generation network (DeepKeyGen) is proposed as a stream cipher generator to generate the private key, which can then be used for encrypting and decrypting of medical images. In DeepKeyGen, the generative adversarial network (GAN) is adopted as the learning network to generate the private key. Furthermore, the transformation domain (that represents the "style" of the private key to be generated) is designed to guide the learning network to realize the private key generation process. The goal of DeepKeyGen is to learn the mapping relationship of how to transfer the initial image to the private key. We evaluate DeepKeyGen using three data sets, namely, the Montgomery County chest X-ray data set, the Ultrasonic Brachial Plexus data set, and the BraTS18 data set. The evaluation findings and security analysis show that the proposed key generation network can achieve a high-level security in generating the private key.

computer science, artificial intelligence, theory & methods,engineering, electrical & electronic, hardware & architecture

What problem does this paper attempt to address?

This paper aims to address the need for medical image encryption to protect patients' private data from being disclosed. Specifically, the paper proposes a deep learning-based key generation network (DeepKeyGen) that serves as a stream cipher generator to produce private keys, which are then used for the encryption and decryption of medical images.

The main contributions include:
1. Proposing a novel deep learning-based private key generator, DeepKeyGen, which utilizes image-to-image translation to achieve the key generation process. The generated private keys have a large key space, high randomness, and sensitivity to changes.
2. Conducting extensive experimental evaluations using three datasets (Montgomery County Chest X-ray Dataset, Ultrasonic Brachial Plexus Dataset, and BraTS18 Dataset) to verify the security and practicality of DeepKeyGen. The results show that the generated private keys have a high level of security and randomness and can effectively encrypt multimodal plaintext medical images. Additionally, DeepKeyGen demonstrates robustness against various known attacks.