CAJ | 학술논문

重音是语言交流中不可或缺的部分，在语言交流中扮演着非常重要的角色。本文基于 ASCCD 朗读语篇语料库，使用MFCC 算法提取每个语音段的融合上下文子段拼接短时谱信息，构建基于 MFCC 算法的上下文短时谱特征集；并选用 NaiveBayes分类器对这类特征集进行建模，而且将具有最大后验概率的类作为该对象所属的类，这种分类方法充分利用了当前语音段的相关语音特性；融合上下文的 MFCC 短时谱特征组在 ASCCD 上能够得到83．6％的汉语重音检测正确率。实验结果证明，融合上下文子段拼接特征规整方法可以用于汉语重音检测研究中。
중음시어언교류중불가혹결적부분，재어언교류중분연착비상중요적각색。본문기우 ASCCD 랑독어편어료고，사용MFCC 산법제취매개어음단적융합상하문자단병접단시보신식，구건기우 MFCC 산법적상하문단시보특정집；병선용 NaiveBayes분류기대저류특정집진행건모，이차장구유최대후험개솔적류작위해대상소속적류，저충분류방법충분이용료당전어음단적상관어음특성；융합상하문적 MFCC 단시보특정조재 ASCCD 상능구득도83．6％적한어중음검측정학솔。실험결과증명，융합상하문자단병접특정규정방법가이용우한어중음검측연구중。
Accent is a critically important component of spoken communication,and plays a very important role in spoken communication.This paper selects from ASCCD corpus and conducts accent by using MFCC algorithm to extract each voice segment of short -time spectrum based on context sub -segment splicing information.After that,the paper builds integration context short -time spectrum feature sets based on MFCC algorithm,and chooses NaiveBayes classifier to model the two fea-ture sets.NaiveBayes is to choose the classes with maximum a posteriori probability as the object's class.This classification method makes full use of the related phonetic features of speech segment.Integration context short -time spectrum of MFCC feature set respectively achieves 83.6% accent detection accuracy on ASCCD.The experimental results indicate that integra-tion context sub -segment splicing feature structured method of MFCC can be used in Chinese accent detection study.