辨证认识话者自动识别系统  

Understand the Automatic Speaker Identification System Dialectically

作　　者：杨俊杰[1,2] 

机构地区：[1]山西大学,山西太原030006 [2]山西警官高等专科学校,山西太原030021

出　　处：《中国司法鉴定》2013年第2期71-75,共5页Chinese Journal of Forensic Sciences

基　　金：2012年度国家社会科学青年基金项目(12CYY015)

摘　　要：为了警示近年来我国司法话者识别领域中出现的一些崇外、盲目追求快速与省事的苗头,结合话者自动识别系统的研究、应用状况,从语音的共性与个性、话者识别结果的相对性与绝对性出发,通过分析比对话者自动识别与语音识别所用的特征参数及实现过程,辨证分析了制约话者自动识别系统准确率的根本原因。指出了话者自动识别系统尚无法达到人们对其的期望,以及适合于司法诉讼领域的话者自动识别系统的发展方向。To warn the abuse of foreign automatic speaker identification system to pursue speed and convenience in recent years,the status of the study and application of automatic speaker identification system was introduced.Based on the commonness and individuality of the voice and the relativity and absoluteness of speaker identification results,this paper compares automatic speaker identification system and automatic speech recognition system in characteristic parameters and process,and analyzes the fundamental reasons of the restricted accuracy of automatic speaker identification system.As a conclusion,the automatic speaker identification system is still unable to achieve what is desired.The development direction of automatic speaker identification in forensic science was put forward.

关 键 词：话者自动识别 语音识别 司法诉讼 特征参数 

分 类 号：DF793.2[政治法律—诉讼法学]