A Hermite Transform based Method for Physiological Signal Watermarking

In this work, a watermarking algorithm for physiological signals is presented. We use photoplethysmography (PPG) signals used to estimate the instantaneous heart rate, and as a watermark, we use fourteen different audio files, emulating the medical audio reports used for diagnosis purposes. We use the Hermite transform (HT) to decompose the PPG signals and insert the watermark in the components with less energy. According to the results, the watermark does not significantly modify the original PPG signals preserving the content. In addition, the extracted watermark is very similar to the original. Finally, the performance of our proposal was assessed using correlation and peak signal-to-noise ratio, and it was compared with other works using physiological signals showing competitive results.