Open Access

ARTICLE

DeepGan-Privacy Preserving of HealthCare System Using DL

Sultan Mesfer Aldossary^*

Prince Sattam Bin Abdulaziz University, Wadi Ad Dawaser, 11990, Saudi Arabia

* Corresponding Author: Sultan Mesfer Aldossary. Email:

Intelligent Automation & Soft Computing 2023, 37(2), 2199-2212. https://doi.org/10.32604/iasc.2023.038243

Received 03 December 2022; Accepted 17 April 2023; Issue published 21 June 2023

Abstract

The challenge of encrypting sensitive information of a medical image in a healthcare system is still one that requires a high level of computing complexity, despite the ongoing development of cryptography. After looking through the previous research, it has become clear that the security issues still need to be looked into further because there is room for expansion in the research field. Recently, neural networks have emerged as a cost-effective and effective optimization strategy in terms of providing security for images. This revelation came about as a result of current developments. Nevertheless, such an implementation is a technique that is expensive to compute and does not handle the huge variety of different assaults that may be made on pictures. The primary objective of the system that has been described is to provide evidence of a complex framework in which deep neural networks have been applied to improve the efficiency of basic encryption techniques. Our research has led to the development and proposal of an enhanced version of methods that have previously been used to encrypt pictures. Instead, the generative adversarial network (GAN), commonly known as GAN, will serve as the learning network that generates the private key. The transformation domain, which reflects the one-of-a-kind fashion of the private key that is to be formed, is also meant to lead the learning network in the process of actually accomplishing the private key creation procedure. This scheme may be utilized to train an excellent Deep Neural Networks (DNN) model while instantaneously maintaining the confidentiality of training medical images. It was tested by the proposed approach DeepGAN on open-source medical datasets, and three sets of data: The Ultrasonic Brachial Plexus, the Montgomery County Chest X-ray, and the BraTS18. The findings indicate that it is successful in maintaining both performance and privacy, and the findings of the assessment and the findings of the security investigation suggest that the development of suitable generation technologies is capable of generating private keys with a high level of security.

---

Keywords

---