Medical Image Encryption Using Distance-Based Biokey Generation

Introduction

Security of medical data and diagnostic images is drawing more attention in the present world. As healthcare evolves by making use of cloud computing and the Internet of Things (IoT), health institutions use third-party servers to store medical data. Guaranteeing the security, privacy, and integrity of these medical data becomes troublesome and is a challenge. Various security algorithms have been investigated over the decades to protect the users’ documents and multimedia contents. When the user wants to store the data on the desktop, laptop, and personalized digital gadgets, there is a login protection mechanism that restricts malicious users to access the data (Panchal et al., 2019). The data can also be protected using file passwords, folder locking, and directory locking methods. The documents can also be protected using security certificates but to be provided and installed by third-party vendors. On the other hand, the encryption algorithms contribute to protecting the documents from malicious access. Personalized documents and multimedia contents of the user are encrypted using encryption algorithms. The key for the encryption algorithms can be generated from various key generation strategies. In this paper, the features of the biometrics traits are used for bio-key generation. The feature selection for bio-key generation is performed using different strategies.

Biometric authentication recognizes a user based on unique physiological attributes (Shaheed et al., 2021). The biometric traits in the form of images and signals, captured using imaging modalities and sensors, deliver various biometric features of the user. The extracted features from the biometrics such as fingerprint, iris, ear, face, and other physiological signals are widely used for user authentication. The above-said biometrics are also be used for key generation to the encryption algorithms. Hence, biometric-based cryptography takes the advantage of biometric-based user authentication and subsequent key generation from the biometric templates. The processing steps ensure two levels of authentication and key generation for encryption and decryption.

Compared to the conventional cryptosystems that use token keys or knowledge keys, a bio-cryptosystem offers a non-repudiable, flexible, and reliable method of security and authentication. The traditional authentication systems have the risk of users forgetting or losing the credentials, while the bio-cryptosystems need not require the user to keep track of any credentials. The security system proposed in this paper makes use of biometrics such as iris and fingerprints for authentication and encryption. A unique key-generation scheme based on distances between minutiae points is applied in this work. Encryption of medical images is done using symmetric key cryptography which enhances faster computation and requires less memory footprint. The robustness of the proposed algorithm is also increased by pixel scrambling. The scrambled version of the plain image at a particular iteration step is selected for encryption using a 256-bit bio-key derived from the fingerprint and iris features of the user. First, the user is authenticated by iris features. The second level of authentication is carried out from fingerprint features. After authentication, the distance between each minutia point with all other minutiae points is evaluated. The distance values are used to decide the bits of the key. The maximum distance decides the length of the key. A key length of 256 bits is obtained by zero padding. A similar bio-key is needed for decryption. Hence, the user needs to execute biometric authentication, bio-key-based decryption, and inverse pixel scrambling to get back the plain image. The performance of the proposed method is also compared with some of the existing image encryption methods.

The paper proceeds as follows. Related works are summarized in section II. The methodology is described in section III. Dataset details and results are presented in section IV which is followed by a conclusion in section V.