Increasing information capacity and improving detection reliability in audio watermarking system

The objective of this paper is to propose a new method for increasing the capacity of watermark information in the new proposed audio watermarking system. The basic idea is to exploit the CDMA (Code Division Multiple Access) technique while considering the multiple embedded watermarks as different users that share the same channel which is the host audio signal. Once the watermarks are embedded, we aim to recover them from only the received signal with low BER (Binary Error Rate) by applying the UISA (Underdetermined Independent Subspaces Analysis) technique. Experimental results prove the performance of the blind watermarking system in terms of inaudibility, detection reliability and robustness against various disturbances.