Abstract
Pulmonary breathing sound plays a key role in the prevention and diagnosis of the lung diseases. Its correlation with pathology and physiology has become an important research topic in the pulmonary acoustics and the clinical medicine. However, it is difficult to fully describe lung sound information with the traditional features because lung sounds are complex and nonstationary signals. And the traditional convolutional neural network cannot also extract the temporal features of the lung sounds. To solve the problem, a lung sound recognition algorithm based on VGGish-BiGRU is proposed on the basis of transfer learning, which combines VGGish network with the bidirectional gated recurrent unit neural network (BiGRU). In the proposed algorithm, VGGish network is pretrained using audio set, and the parameters are transferred to VGGish network layer of the target network. The temporal features of the lung sounds are extracted through retraining BiGRU network with the lung sound data. During retraining BiGRU network, the parameters in VGGish layers are frozen, and the parameters of BiGRU network are fine-tuned. The experimental results show that the proposed algorithm effectively improves the recognition accuracy of the lung sounds in contrast with the state-of-the-art algorithms, especially the recognition accuracy of asthma.
Highlights
Pulmonary auscultation is one of the effective methods to diagnose the lung diseases
From the view of the spatial domain and the time domain, a lung sound recognition model based on VGGish-bidirectional gated recurrent unit neural network (BiGRU) is proposed by combining VGGish convolutional neural network with BiGRU recurrent neural network
The experiments consist of six parts: the effectiveness verification of the model, the effect of the heart sounds on results, the influence of different time-frequency analysis methods on results, the effect of transfer learning on results, the influence of the retraining layer on results, and comparison of different methods
Summary
Pulmonary auscultation is one of the effective methods to diagnose the lung diseases. The method extracted MFCC features, and a two-layer convolutional neural network (2L-CNN) was used to train and recognize the lung sounds. MFCC features were utilized to identify the lung sounds by a five-layer convolutional neural network (5L-CNN), and better recognition results were obtained in [25]. Short time Fourier transform was used to analyze time-frequency features of the lung sounds and the lung sounds were classified into three categories by combining two-layer convolutional neural networks with two-layer full connections in [26]. The temporal feature of the lung sound signals is captured by taking the bidirectional gated recurrent unit neural network (BiGRU) as the retraining layer of transfer learning. It can be seen from the results that the low frequency noise and the heart sound components are effectively deleted
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
