An efficient watermarking algorithm for digital audio data in security applications

Area of Science:

• Digital Signal Processing
• Embedded Systems
• Information Security

Background:

• Transform-domain audio watermarking offers robustness but suffers from high computational costs, limiting real-time applications.
• Existing methods are often too slow for time-critical security needs, especially with long audio signals.

Purpose of the Study:

• To develop a computationally efficient and fast audio watermarking system for real-time security applications.
• To leverage parallel processing on embedded systems to overcome the speed limitations of transform-domain watermarking.

Main Methods:

• A hybrid transform domain using fractional Charlier transform (FrCT) and dual-tree complex wavelet transform (DTCWT) was employed.
• Intensive watermarking steps were parallelized and implemented on a cluster of four Raspberry Pi 4B devices using Message Passing Interface (MPI).

Main Results:

• The parallelized system achieved significant temporal improvements: approximately 70% with 4 cores, 80% with 8 cores, and 90% with 16 cores.
• The system demonstrated robustness against manipulation, high payload capacity, and high imperceptibility.

Conclusions:

• The proposed parallel audio watermarking system effectively addresses the computational cost challenge.
• The portable and low-power Raspberry Pi cluster enables efficient, real-time watermarking for smart city applications.