Biometric based audio ownership verification using discrete wavelet transform and SVD techniques

The digital revolution in audio industry has brought new ownership verification challenges, as a recording can be copied and distributed easily. Audio watermarking has been proposed as an effective solution to solve or mitigate the impact of these challenges. This paper investigates a previously-reported digital audio watermarking algorithm with the aim of improving its features and performance. The modified algorithm is then used to develop an audio ownership verification system. This is achieved by comgining the advantages of Discrete Wavelet Transform (DWT) and Singular Value Decomposition (SVD) techniques along with utilizing biometric features (thumbprint) of the owner as a digital watermark. A fully functional ownership verification system is developed and its performance is assessed experimentally. When compared to the original algorithm, the suggested system show significant improvements in terms of minimizing the required minimum period of original audio signal, improving quality of watermarked audio signal, and detecting the applied attacks.