Multi-channel lung sound classification with convolutional recurrent neural networks
Multi-channel lung sound classification with convolutional recurrent neural networks
复制标题
DOI：
10.1016/j.compbiomed.2020.103831
复制
发表时间：
2020-07-01
影响因子：
7.7
通讯作者：
Pernkopf, Franz
中科院分区：
文献类型：
--
作者：
Messner, Elmar;Fediuk, Melanie;Pernkopf, Franz
本文提出了一种利用频谱、时间和空间信息进行多通道肺声分类的方法。特别是，我们提出了一种基于帧的分类框架，以卷积递归神经网络处理多通道肺录音的完整呼吸周期。利用我们最近开发的16通道肺录音设备，我们在临床试验中收集肺健康受试者和特发性肺纤维化（IPF）患者的肺录音。从肺录音中提取频谱特征，并比较不同的深度神经网络结构进行二分类，即健康与病理。我们提出的卷积递归神经网络分类框架优于其他网络，其F-1得分接近92%。结合我们的多通道肺音记录装置，我们提出了一种多通道肺音分析的整体方法。
In this paper, we present an approach for multi-channel lung sound classification, exploiting spectral, temporal and spatial information. In particular, we propose a frame-wise classification framework to process full breathing cycles of multi-channel lung sound recordings with a convolutional recurrent neural network. With our recently developed 16-channel lung sound recording device, we collect lung sound recordings from lung-healthy subjects and patients with idiopathic pulmonary fibrosis (IPF), within a clinical trial. From the lung sound recordings, we extract spectrogram features and compare different deep neural network architectures for binary classification, i.e. healthy vs. pathological. Our proposed classification framework with the convolutional recurrent neural network outperforms the other networks by achieving an F-score of F-1 approximate to 92%. Together with our multi-channel lung sound recording device, we present a holistic approach to multi-channel lung sound analysis.