阈值自适应有声出版物语音自动分割算法  

Speech automatic segmentation algorithm of audio publication with adaptive threshold adjustment

作　　者：张俊星[1] 石立新[1] 王都生[1] 

机构地区：[1]大连民族学院机电信息工程学院,辽宁大连116600

出　　处：《计算机应用》2010年第2期567-570,共4页journal of Computer Applications

摘　　要：为完成有声出版物中的语音自动分割,建立了一种时间阈值自适应加相似度判决的系统分割模型。时间阈值的确定是系统设计中的一个难点,为此基于脚本中的先验知识提出了时间阈值自适应分割算法。为提高系统的抗干扰能力以增强其适用性,提出了基于语音单元相似性进行结果验证的新方法。测试表明录音过程中不同语音单元间略作停顿时,机器分割率在95%以上,分割的正确率100%。In order to achieve the automatic speech segmentation of audio publication, the system model with adaptive time threshold adjustment and similarity detection was established. The adaptive threshold adjustment algorithm was put forward to set the time threshold based on the prior knowledge of script. In order to improve the anti-interference and adaptive abilities, a new algorithm based on the similarity detection of speech units was provided. The test results show that the accuracy of segmentation is 100% when the system has no interference, but the accuracy of segmentation is 98.8% when two interference signals are added to each speech files on average. Practical application shows that the system can meet the automatic segmentation requirements.

关 键 词：有声出版物 语音分割 时间阈值自适应 相似性分析 

分 类 号：TP391.42[自动化与计算机技术—计算机应用技术] TN912.34[自动化与计算机技术—计算机科学与技术]