摘要
为了探究短时语音在法庭说话人识别中的应用价值,利用基于似然比框架的法庭自动说话人识别系统,对典型案件条件下的短时语音进行了法庭说话人识别的验证测试和分析比较。通过对不同时长、不同校准集人数规模及其音频数量的测试比较,量化评估了系统在短时语音条件下的识别性能、3种因素对识别性能的影响以及短时语音在司法实践中的应用价值。研究结果表明,短时语音条件下系统仍具有良好的准确性和可靠性,这不仅验证了该系统的有效性和顽健性,也说明了短时语音在司法实践中的应用潜力。
To investigate the application value of short-duration speech in forensic speaker recognition,a forensic automatic speaker recognition system based on the likelihood ratio framework is employed to conduct validation tests and comparative analysis on short-duration speech under typical case conditions.The recognition performance of the system under short-duration speech conditions,the impact of three factors on the recognition performance,and the application value of short-duration speech in forensic practice are quantitatively evaluated by comparative testing of varying durations,different calibration set size of the number of speakers and their audio.The results show that the forensic automatic speaker recognition system still has good accuracy and reliability under short-duration speech conditions.This not only validates the effectiveness and robustness of the system,but also demonstrates the application potential of short-duration speech.
作者
张翠玲
刘明星
ZHANG Cuiling;LIU Mingxing(School of Criminal Investigation,Southwest University of Political Science and Law,Chongqing 401120,China;Institute of Intelligent Justice,Southwest University of Political Science and Law,Chongqing 401120,China;Chongqing Institutes of Higher Education Forensic Science Key Laboratory,Chongqing 401120,China)
出处
《中国人民公安大学学报(自然科学版)》
2025年第2期100-108,共9页
Journal of People’s Public Security University of China(Science and Technology)
基金
重庆市自然科学基金创新发展联合基金(市教委)A类项目(CSTB2022NSCQ-LZX0007)
西南政法大学校级重点委托项目(023-XZWTZD06)。
关键词
短时语音
法庭说话人识别
自动说话人识别
似然比
short-duration speech
forensic speaker recognition
automatic speaker recognition
likelihood ratio
