Computer Engineering and Applications ›› 2009, Vol. 45 ›› Issue (12): 131-133.DOI: 10.3778/j.issn.1002-8331.2009.12.043

• 数据库、信号与信息处理 • Previous Articles     Next Articles

New effective method on content based audio feature extraction

ZHENG Ji-ming1,WEI Guo-hua2,WU Yu2   

  1. 1.Institute of Applied Mathematics,Chongqing University of Posts and Telecommunications,Chongqing 400065,China
    2.Institute of Computer Science and Technology,Chongqing University of Posts and Telecommunications,Chongqing 400065,China
  • Received:2008-03-03 Revised:2008-06-06 Online:2009-04-21 Published:2009-04-21
  • Contact: ZHENG Ji-ming

有效的基于内容的音频特征提取方法

郑继明1,魏国华2,吴 渝2   

  1. 1.重庆邮电大学 应用数学研究所,重庆 400065
    2.重庆邮电大学 计算机科学与技术学院,重庆 400065
  • 通讯作者: 郑继明

Abstract: Feature extraction is the foundation of the audio classification,and good features will enhance the classification accuracy effectively.In this paper,Mel-frequency cepstrum coefficients are extracted from frequency domain of audio.At the same time,features are extracted from wavelet domain after discrete wavelet transform is done for each frame of the audio.Then the features from the frequency domain and wavelet domain are combined to calculate the statistical features.Finally,audio template is established according to the Support Vector Machine(SVM),and it is classified and identified into speech,music and speech with music.Tests show that the method gets comparatively high identification accuracy.

摘要: 音频特征提取是音频分类的基础,好的特征将会有效提高分类精度。在提取频域特征Mel频率倒谱系数(MFCC)的同时,对每一帧信号做离散小波变换,提取小波域特征,把频域和小波域特征相结合计算其统计特征。通过SVM模型建立音频模板,对纯语音、音乐及带背景音乐的语音进行分类识别,取得了较高的识别精度。