针对语音变换的语音篡改检测

Speech Tampering Detection for Voice Transformation

针对使用语音变换技术的语音篡改,提出一种自动检测方法。在分析语音变换基本模型和变换语音失真的基础上,提取语音信号的声道参数以及相关的信号统计量,并通过支持向量机递归特征消除法,选择出对语音变换比较敏感的特征作为分类特征,使用支持向量机进行语音变换检测和变换语音的说话人性别判别。对于一种语音变换软件的实验结果表明,该方法具有较高的检测准确率,其中语音变换检测的平均准确率为94.90%,变换语音的说话人性别判别平均准确率为92.09%。

An automatic speech tampering detection method for voice transformation is pro- posed. Based on the analysis of the basic model of voice transformation and distortions intro- duced by it, vocal tract parameters and statistics of speech signals are extracted. Sensitive fea- tures are selected as classifying features by support vector machine recursive feature elimination method. Support vector machine classifiers are applied to get the decision of voice transforma- tion detection and speaker gender of transformed speech. Experimental results on a kind of voice transformation software show that the proposed method has high detection accuracy, where the average accuracy is 94.90% for voice transformation and 92.09% for speaker gender detection of transformed speech.