New Zero-Watermarking Algorithm Using Hurst Exponent for Protection of Privacy in Telemedicine

Abstract

Telemedicine has numerous potential applications in the medical field due to the significant progress of telecommunication and information technology in recent years. In any category of telemedicine, such as offline, remote monitoring, and interactive, medical data and personal information of an individual must be transmitted to the healthcare center. An unauthorized access to the data and information is unacceptable in a telemedicine application, because it may create unavoidable circumstances for a person’s private and professional life. To avoid any potential threat of identity exposure in telemedicine, a zero-watermarking algorithm to protect the privacy of an individual is proposed in this paper. The proposed algorithm embeds the identity of a person without introducing any distortion in medical speech signals. Two measures, namely, Hurst exponent and zero-crossing, are computed to determine the suitable locations in the signal for insertion of identity. An analysis of the signals indicates that unvoiced speech frames are reliable in insertion and extraction of identity, as well as robust against a noise attack. In the proposed zero-watermarking algorithm, identity is inserted in a secret key instead of a signal by using a 1-D local binary operator. Therefore, imperceptibility is naturally achieved. Experiments are performed by using a publicly available voice disorder database, and experimental results are satisfactory and show that the proposed algorithm can be reliably used in telemedicine applications.