Detection of acoustic features in human biological sounds based on statistical approach

在这项研究中，我们设计了一种新的分类程序，用于区分正常和异常的呼吸音的基础上随机的方法。我们的程序的主要特点是，两个随机模型被用来精确地检测异常呼吸音：隐马尔可夫模型的声学频谱特征和二元模型的发生在每个呼吸周期的声学段。我们还设计了一种方法来分类的情感集群使用韵律特征。在我们的方法中，我们使用特定的声学段的持续时间比共振哭段和沉默段在婴儿的哭声作为韵律特征。我们的方法使用段持续时间比的分类性能显着优于使用频谱特征的方法。

In this research project, we devised a novel classification procedure for distinguishing between normal and abnormal respiratory sounds on the basis of stochastic approach. The main characteristic of our procedure is that two stochastic models are used to detect abnormal respiratory sounds precisely: hidden Markov models for acoustic spectral features and bigram models for the occurrence of acoustic segments in each respiratory period. We also devised an approach to the classification of emotion clusters using prosodic features. In our approach, we use the duration ratios of specific acoustic segments-resonant cry segments and silence segments-in the infants' cries as prosodic features. The classification performance of our approach using the segment duration ratios was significantly better than that of the method using spectral features.