在似然比证据评估体系下研究元音的共振峰特征在法庭语音证据强度评估中的应用,提供了似然比计算方法,并在包含42人语音的数据库中对法庭说话人识别系统的性能及可靠性进行了检验。结果表明:使用共振峰特征的法庭说话人识别系统,在只使用一个元音的情况下具有良好的识别率。该方法不仅能正确识别说话人,而且能根据当前语音证据的差异,量化该语音样本作为证据的力度,为法庭提供科学合理的证据评估结果。 The application of the formant feature of vowel to the assessment of court voice evidence intensity under the likelihood ratio evidence evaluation system provides the likelihood ratio calculation method and the performance of courtroom speaker recognition system in a database of 42 voices And the reliability has been tested. The results show that the forensic speaker recognition system using formant features has a good recognition rate with only one vowel. The method not only can correctly identify the speaker, but also can quantify the speech sample as the evidence strength according to the difference of the current speech evidence and provide the court with scientific and reasonable evidence evaluation results.