摘要
音频分类是提取音频结构和内容语义的重要手段,是基于内容的音频、视频检索和分析的基础。支持向量机(SVM)是一种有效的统计学习方法。本文提出了一种基于SVM的音频分类算法。将环境音分为6类:车鸣声,钟声,风声,冰块声,机床声和雨声。特征抽取是音频分类的基础。本文从帧层次上深入分析了不同类音频之间的区别性特征,包括频域能量,子带能量,过零率,频率中心,带宽,基音频率及MFCC(Mel-Frequency Cepstral Coefficients)。实验结果表明,支持向量机模型的环境音分类性能较好,最优分类精度达到97.73%。
Audio classification is an important access to extract audio structure and content, and is a basis for further audio/video retrieval and analysis. Support vector machines （SVM） is a valid statistic learning method. In this paper, the work on audio classification based on SVM is presented. Six environmental audio classes are considered in this paper： the sound of vehicle, bell, wind, ice, machine tool and rain. Feature extraction is the foundation of audio classification. Audio features are analysed deeply in frame level, including frequency energy, sub-band energy, ZCR, frequency centroid, bandwidth, pitch frequency and MFCC （Mel-Frequency Cepstral Coefficients）. The experimental results show that SVM is excellent for environmental audio classification, and the optimal classification accuracy is up to 97.73%.
出处
《电子测量技术》
2008年第9期121-123,132,共4页
Electronic Measurement Technology
关键词
环境音分类
支持向量机
MFCC
environmental audio classification
support vector machine
MFCC
