摘要
本文提出了一种新的基于人工神经网络的音频水印算法。该算法利用人工神经网络来智能地估计原宿主信号的水印强度因子,并以这些强度因子的值作为确定水印嵌入强度的依据,以此种方式嵌入水印后的信号其能量频谱能够总是保持在宿主信号的最小听觉掩蔽门限之下。实验结果表明,在音频作品里面以这样的方式加入水印,不但能够保持相当好的透明性,而且对诸如加噪、重采样、重新量化等攻击表现出较好的稳健性。
A novel scheme for watermarking on audio signals using artificial neural networks(ANNs) is presented in this paper. The ANN is used to estimate the watermark scaling factor(WSF) intelligently from the knowledge of host audio signal. The watermarked signal" s power spectrum remains below the minimum masking threshold (MMT) of the host signal when these WSFs are used in the watermarking process. Experimental results show that it not only ensures inaudibility of the watermarked signal but also is robust to familiar attacks.
出处
《微计算机信息》
北大核心
2005年第11Z期184-186,共3页
Control & Automation
基金
上海工程技术大学青年基金
编号:2004Q16
关键词
人工神经网络
水印
DCT变换
artificial neural network (ANN)
watermarking
discrete cosine transform