Environmental Sound Classification using Hybrid SVM/KNN Classifier and MPEG-7 Audio Low-Level Descriptor
Environmental Sound Classification using Hybrid SVM/KNN Classifier and MPEG-7 Audio Low-Level Descriptor
复制标题
DOI：
10.1109/ijcnn.2006.246644
复制
发表时间：
2006-07
期刊：
影响因子：
--
通讯作者：
Jia-Ching Wang;Jhing-Fa Wang;Wai-He Kuok;Cheng-Shu Hsu
中科院分区：
文献类型：
--
作者：
Jia-Ching Wang;Jhing-Fa Wang;Wai-He Kuok;Cheng-Shu Hsu
在本文中，我们提出了一个新的环境声音分类架构。建议的声音分类器在帧级执行，并融合支持向量机（SVM）和k近邻规则（KNN）。在特征选择中，三个MPEG-7音频低级别描述符，频谱质心，频谱扩展，频谱平坦度被用作声音特征，以发挥其在声音分类的能力。在一个12类的声音数据库上进行的实验可以达到85.1%的准确率。与基于音频频谱投影特征的HMM声音分类器的性能比较表明了该方法的优越性。
In this paper, we present a new environmental sound classification architecture. The proposed sound classifier is performed in frame level and fuses the support vector machine (SVM) and the k nearest neighbor rule (KNN). In feature selection, three MPEG-7 audio low-level descriptors, spectrum centroid, spectrum spread, and spectrum flatness are used as the sound features to exploit their ability in sound classification. Experiments carried out on a 12-class sound database can achieve an 85.1 % accuracy rate. The performance comparison between the HMM sound classifier using audio spectrum projection features demonstrates the superiority of the proposed scheme.