中国民族乐器的特征值提取和分类

文章处理的是中国民族乐器的识别分类问题.提出了一种基于合适的音频特征值选择方法,该方法在基于MPEG_7标准的声学特征,在特征值的时间特性上进行改造,并加入改进后的特征值.从提取的音频特征值数据集选择K-最近邻算法.特别是对没有加入新特征值和加入新特征值后生成分类器模型的性能进行了比较.实验结果证明新特征值的加入提高了分类器的F1度量值.

Audio Feature Extraction and Classification of the Chinese National Musical Instrument

This paper address the problem of classifying database of Chinese national musical instrument audio.An efficient algorithm is proposed for selecting the most appropriate signal features for a given classification task.The method is based on MPEG-7 standards in acoustic feature on the time property of transformation,and join the improved values.The classifiers KNN,classifiers are compared for the task of database classifiacation.Especially compared with created classifier model performance not to join the new characteristic value and adding new features.The experimental results show that the F1 measure is obtained when the novel signal features are used.