| 一种基于支持向量机的数字音频认证方法 |
| |
| 引用本文： | 林晓丹.一种基于支持向量机的数字音频认证方法[J].华侨大学学报(自然科学版),2011,32(2):153-155. |
| |
| 作者姓名： | 林晓丹 |
| |
| 作者单位： | 华侨大学,信息科学与工程学院,福建,泉州,362021 |
| |
| 基金项目： | 华侨大学科研基金资助项目(09HZR12) |
| |
| 摘 要： | 提出一种基于数字水印技术的音频内容认证方法.选取稳定的梅尔倒谱特征构造特征向量,利用支持向量机自适应地选取合适的帧进行水印的嵌入和提取.结果表明,算法在保证嵌入水印的不可感知性的同时,能够有效地区分恶意的内容篡改和非恶意的常规信号处理操作,准确地定位篡改发生的位置.
|
| 关 键 词： | 数字音频 水印技术 认证 梅尔倒谱系数 支持向量机 |
An SVM-Based Digital Audio Authentication Method |
| |
| LIN Xiao-dan.An SVM-Based Digital Audio Authentication Method[J].Journal of Huaqiao University(Natural Science),2011,32(2):153-155. |
| |
| Authors: | LIN Xiao-dan |
| |
| Institution: | LIN Xiao-dan (College of Information Science and Engineering,Huaqiao University,Quanzhou 362021,China) |
| |
| Abstract: | A watermarking-based method for audio content authentication is proposed.Mel frequency cepstral coefficient(MFCC) is adopted to construct training vectors for support vector machine(SVM).Adaptive watermark embedding and extraction is achieved by the well-trained SVM.Experimental results demonstrate that this approach not only can distinguish malicious tampering from content-preserving operations,but also accurately locate regions that have undergone malicious manipulations while keeping the inaudibility of ... |
| |
| Keywords: | digital audio watermarking authentication Mel frequent cepstral coefficient support vector machines |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
