盐城工学院学报（自然科学版）
鹽城工學院學報（自然科學版）
염성공학원학보（자연과학판）
Journal of Yancheng Institute of Technology(Natural Science Edition)
2015年
3期
44-46
,共3页
端点检测%低信噪比%频带方差
耑點檢測%低信譟比%頻帶方差
단점검측%저신조비%빈대방차
endpoint detection%low SNR%frequency band variance
语音识别中语音信号端点检测的准确度起着至关重要的作用,传统的频带方差算法在强背景噪声环境下性能下降.提出一种改进算法,在保证算法复杂度较低的条件下,大大提高了检测的准确度.仿真证明,该算法在低信噪比条件下也具有较强的鲁棒性.
語音識彆中語音信號耑點檢測的準確度起著至關重要的作用,傳統的頻帶方差算法在彊揹景譟聲環境下性能下降.提齣一種改進算法,在保證算法複雜度較低的條件下,大大提高瞭檢測的準確度.倣真證明,該算法在低信譟比條件下也具有較彊的魯棒性.
어음식별중어음신호단점검측적준학도기착지관중요적작용,전통적빈대방차산법재강배경조성배경하성능하강.제출일충개진산법,재보증산법복잡도교저적조건하,대대제고료검측적준학도.방진증명,해산법재저신조비조건하야구유교강적로봉성.
The accuracy of speech recognition in speech signal endpoint detection plays an important role.Traditional frequency band variance algorithm gives a bad performance in strong noise environment.In ensuring a lower computational complexity con-ditions,Put forward an improved algorithm,which can greatly improve the accuracy of detection.The simulation results show that the algorithm has strong robustness under low SNR condition.
