基于变换域的数字音频水印技术研究-电路与系统专业论文.docxVIP

水印后的音频作比较分析。实验表明采用小波包分解音频并进 水印后的音频作比较分析。实验表明采用小波包分解音频并进 行能量特征提取，能有效地实现声音水印的检测。 4)提出了一种基于改进后的小波变换的抗剪切音频水印算法。该 算法最大的优点是利用水印序号的嵌入来弥补剪切攻击所造 成的水印信息的丢失。实验表明，每嵌入一个水印元素之前嵌 入一组水印序号的二进制序列组，能有效地降低剪切攻击损 害。 关键词：数字音频，水印，鲁棒性，变换域，同步攻击 Ⅱ ABSTRACTDigital ABSTRACT Digital audio watermarking technology is an important branch of digital watermarking technology，and it has been applied and valued widely in the fields of copyright protection， private voice communication．etc，and has a significance to the development of future digitization．Based ON the existing theories， aiming at researching digital audio watermarking technology in the transform domain，three digital audio watermarking algorithms are showed．What paper do and research are the fol lowing respects： 1)Several typical of audio watermarking algorithms at home and abroad are introduced and analyzed，then summari ze and compare their di fferences． 2)A adaptive double digital audio watermarking algorithm based on chaos theory in complex cepstrum i s proposed． A string of chaotic sequence produced by the Logistic chaos mapping is changed to a binary image，and then XOR with the primitive watermark image，SO that the watermark i s encrypted，and the securi ty wi l l be enhanced．．In addition，with two different pictures as the watermark， embedding process needs do twi ce，that i s，the second M embedding embedding process is a kind of attack for the first embedding process．The results show that this algorithm can show a preferable imperceptible and robustness． 3)A speech watermarking algorithm based on discrete cosine transform and a watermark detection algorithm based on wavelet packet decomposition are proposed． With a range of voice as a watermark，the primitive audio is transformed by DCT，then the watermark is embedded in the low-frequency coefficients that contain high energy． Based on wavelet packet decomposition of feature extract ion method，a pi ece of audio added by Gauss whi te noise is used to compare with the primitive audio and