Symmetric Cryptography With Shift 2n-1, Hash Transformation, Optimization-Based Controller for Medical Image Infosecurity: Case Study in Mammographic Image

Pi-Yun Chen,Jian-Xing Wu,Chao-Lin Kuo,Neng-Sheng Pai,Chien-Ming Li,Chia-Hung Lin

doi:10.1109/jphot.2020.2987769

Abstract

Among imaging examinations, mammography noninvasively aids in the early detection and diagnosis of diseases inside the body, such as lung cancer or breast tumor, for first-line examination and helps physicians to treat medical conditions. As digital medical images can be transmitted by computer or telecommunication network, ensuring image infosecurity in public spaces is difficult. In infosecurity considerations, digital signature and encryption of individual anamnesis plain images should prevent data from being stolen, tampered, or peeked by unauthorized people. Hence, infosecurity has become an important issue in the digital era. This study proposes right / left shift 2n−1 pixel values, hash transformation with multi-secret gold keys, and an optimization-based controller to integrate two symmetric cryptography processes for medical image encryption and decryption with the same secret gold keys. Then, swarm optimization-based controllers, which are based on particle swarm optimization (PSO) and firefly algorithm (FA), are implemented to search the decryption key parameters. For a case study on mammographic images consisting of medical X-ray images of 50 benign tumors and 50 malignant tumors, the peak signal-to-noise ratio is used to evaluate the relation between the original plain images and decrypted images. Conclusively, the FA -based controller performed better than the PSO -based controller in terms of recovery reliability and speeds up the computation of the desired optimal solution.

Highlights

With the digitalized image data and popularization of network communications, data protection is becoming more important via wired or wireless communication in public spaces
For 100 digital X-ray images, including case studies of benign and malignant tumors, the peak signal-to-noise ratio (PSNR) [30], [31] is used to evaluate the performance of proposed symmetric cryptographic methods; the results indicate that the recoverable image is reliable for further diagnostic applications and decision making
A symmetric cryptographic method based on shift cipher, hash transformation, and swarm optimization-based controller was proposed for digital X-ray images in this study

Summary

Introduction

With the digitalized image data and popularization of network communications, data protection is becoming more important via wired or wireless communication in public spaces. An attacker could access these images to change the outcome or evidence of the patient’s diagnosis, such as aneurysms, heart disease, and tumors / cancers (breast cancers), and deceive radiologists This examination showed that medical images are tampered by using advanced artificial intelligent methods, resulting in misjudgment by radiologists. Adversarial attacks could cause a deep neural network classifier to change its predictions, resulting in incorrect predictions [2]–[4] These image perturbations could cheat multiple machine learning methods. Information security researchers have pointed out that detecting tampered medical images is difficult if they have not undergone “digital signature” and “data encryption.”. Information security researchers have pointed out that detecting tampered medical images is difficult if they have not undergone “digital signature” and “data encryption.” In view of this condition, corresponding information security and protection measures should be obtained

Methods

Results

Conclusion