An efficient scheme for secure domain medical image fusion over cloud

Abstract

The exponential growth in the medical images is making the healthcare industry move towards cloud-based paradigm, which has vast storage and high end processing facilities. However, moving medical images containing highly sensitive data to third-party cloud servers brings in serious security threats. Even though encrypting medical images before outsourcing using traditional encryption schemes seem to be a feasible solution, that can not support encrypted domain processing. In this paper, we propose an affine Hill cipher based scheme for encrypted domain medical image fusion. The random vectors used in this scheme are carefully designed to preserve the randomness and security properties when operations are performed on the encrypted data. The proposed scheme offers data privacy and supports encrypted domain processing with no additional storage burden at the cloud side and very low computational burden at the healthcare provider side. The security of the proposed scheme is evaluated through extensive cryptanalysis in terms of resistance against various statistical attacks. The performance of the proposed scheme is analyzed by comparing various metrics of encrypted domain MR-CT/PET image fusion results with those of plaintext domain fusion. The values of structural similarity index, normalized correlation coefficient and structural content are 1 and the image quality index is 0.999, which show that the proposed encrypted domain image fusion provides same accuracy levels as that of plaintext domain image fusion.