Reversible Data Hiding for Electronic Patient Information Security for Telemedicine Applications

摘要

Cloud computing along with the Internet of Things (IoT) is proving to be an essential tool for delivering better healthcare services. However, maintenance, privacy, confidentiality, and security of Electronic Health Information (EHI) pose a huge challenge in telemedicine. The sharing of EHI with a remote doctor over the cloud is an important issue since the minute variation may lead to the wrong diagnosis. Despite the plethora of research in this field, there is an immense necessity to develop the algorithms for enhancing security in e-healthcare systems. In this paper, an innovative Reversible Data Hiding (RDH) scheme using Lagrange’s interpolation polynomial, secret sharing, and bit substitution for EHI security has been proposed. The cover medical image is sub-sampled, into four shares. Image interpolation is used to enlarge the subsamples, for hiding EHI. The secret information is processed using Lagrange’s interpolation polynomial before being embedded in the various cover image shares. The data is embedded into the interpolated sub-sampled shares at the locations pre-defined by the algorithm. The distributive nature of embedded data enhances the security of the proposed framework while maintaining reversibility. We show that only 75% of shares are required to obtain the whole embedded data and the undistorted cover image. The proposed scheme outperforms the schemes under comparison in terms of imperceptibility and payload. It can reversibly embed 163,840 bits (0.75 bits per pixel) with an average PSNR of about 52.38 dB. The average values of relative entropy, the difference in relative entropy, standard deviation, and cross-correlation are 7.3242, 0.0382, 65.0539, and 0.9838, respectively, for the first sub-sample. It shows an increase of about 3 dB for a payload of 1, 30,000 bits when compared to the state-of-art. Further, it is pertinent to mention that the proposed scheme has lower computational complexity and is hence useful for e-healthcare applications. Given all the attributes of the scheme along with its lower computational complexity, it is suitable for EHI security in a distributive environment like cloud computing.