广西师范学院学报(自然科学版)
廣西師範學院學報(自然科學版)
엄서사범학원학보(자연과학판)
JOURNAL OF GUANGXI TEACHERS COLLEGE(NATURAL SCIENCE EDITION)
2012年
4期
88-91
,共4页
语音识别%DTW%人工神经网络%拒识
語音識彆%DTW%人工神經網絡%拒識
어음식별%DTW%인공신경망락%거식
speech recognition%DTW%Artificial Neural Networks%rejection
语音识别是当今信息领域标志性技术,是最具发展前景的综合性学科之一.该文用语音识别系统模式匹配原理,对比分析了模板匹配算法DTW与人工神经网络中的ANN算法在语音识别据识中的应用;提出了一种基于单个神经元的拒识算法,实验结果表明该算法能达到较高的拒识率.
語音識彆是噹今信息領域標誌性技術,是最具髮展前景的綜閤性學科之一.該文用語音識彆繫統模式匹配原理,對比分析瞭模闆匹配算法DTW與人工神經網絡中的ANN算法在語音識彆據識中的應用;提齣瞭一種基于單箇神經元的拒識算法,實驗結果錶明該算法能達到較高的拒識率.
어음식별시당금신식영역표지성기술,시최구발전전경적종합성학과지일.해문용어음식별계통모식필배원리,대비분석료모판필배산법DTW여인공신경망락중적ANN산법재어음식별거식중적응용;제출료일충기우단개신경원적거식산법,실험결과표명해산법능체도교고적거식솔.
@@@@Speech recognition is a sign of information technology and is one of the most promising com-prehensive disciplines .This paper briefly describes the speech recognition system and its pattern matching theory .Comparative analysis is made of the template matching algorithms DTW and artificial neural net-work algorithm ANN for rejection of speech recognition system .A single neuron rejection algorithm is pro-posed .The results shows that the algorithms achieve a high rejection rate .