Computationally efficient image encryption technique based on selective pixel diffusion

Abstract

Medical images play a vital role in disease diagnosis. When the medical images are communicated through an insecure transmission channel, their chances of being accessed by an unauthorized user increase resulting in the loss of patients' sensitive personal data. Thus, providing security to such image data while transmitting it over an insecure communication network becomes crucial. This work presents a computationally efficient cryptosystem for encrypting medical images. The encryption process consists of multiple phases. In the first phase the control parameters and initial values for the various chaotic maps used, are evaluated. This phase is followed by encryption in which these evaluated values are used to obtain the chaotic sequences for encryption. In the subsequent phases, we make use of a new approach of selective, pixel-dependent diffusion to obtain the cipher image. The effectiveness of our cryptosystem is evaluated using security analysis and execution time analysis. The obtained outcome shows a high-security level compared to already existing state-of-the-art techniques. In addition, the computational complexity of our scheme is very small (0.1sec for encrypting a 256×256 image) making it suitable for real-time smart health applications.