Consideration of improvement of the precision of the method which estimates the detection error rates in a certain digital watermarking in audio signal

摘要

Since the watermarked digital contents can be altered by various signal processing, it is difficult in digital watermarking to detect the embeded watermarks perfectly. Then, we have researched on estimating the detection error rates for a digital watermarking, and proposed the estimating method for a certain digital watermarking for audio signal. In this paper, we improve the pricision of the method by selecting data which are used for estimation out of those which are made redundantly from inputs of the method. We evaluated the proposed method using MPEG 1 audio layer III as a signal processing. We show that the estimation is more accurately than before.