A reversible and multipurpose ECG data hiding technique for telemedicine applications

Biomedical signals serves as a backbone of telemedicine industry and its analysis help the remote diagnosis possible. Security of biomedical data is one of the critical issues during its transmission from one storage device to another storage device. In this work, a completely reversible data hiding method is proposed for the ECG (Electrocardiogram) data. The proposed scheme is designed in such a way that it can find out false ownership claims as well as detect the tampered region of ECG data. In addition, original ECG signal is perfectly reconstructed from watermarked signal. The reversibility is tested for a large set of 460 different ECG signals generated from MIT-BIH arrhythmia database and complete reversibility (100%) is obtained each time. In proposed approach, deep neural network is used for an improved error prediction and prediction error expansion (PEE) is combined with it to ensure reversibility. Proposed method is a high capacity method and data hiding of 0.99 bps (bits per sample) is achieved with very little distortion in watermarked signal. The performance of proposed work is evaluated through percentage residual difference (PRD), signal to noise ratio (SNR) and normalized cross correlation (NCC). The most important contribution of proposed work is its multipurpose nature: ownership detection, tamper localization and 100% reversibility. Tamper detection and localization performance of the proposed method is also tested against different attacks and it is found to be quite good. (C) 2019 Elsevier B.V. All rights reserved.