In this study, a method based on deep learning has been proposed for the classification of lung sounds. For this purpose, the Convolutional Neural Network (CNN) has been designed. In addition, experiments are carried out using different machine learning methods based on feature extraction.
Figure A. Schema of the Proposed MethodPurpose: The focuses is on automatic diagnosis of lung diseases, one of the most important issues in public health. There have been many studies on this subject in the literature, but most of these studies consist of traditional methods. The aim of this study is to increase the classification performance of lung sounds with deep learning.
Theory and Methods:The proposed method for the classification of lung sounds consists of five steps. First, lung sound signals are pre-processed, and spectrograms are obtained. After applying the data augment process to spectrograms, spectrogram images are given as input to the designed ESA model and the classification process is made.Results: Experiments to evaluate the effectiveness of different methods are carried out using the ICBHI 2017 data set consisting of four classes commonly used in the literature. On average, 64.5% accuracy was obtained from the proposed method. The simulation and experimental results are presented and compared in Section 6.
Conclusion:In this study, deep learning is discussed to improve the classification performance of lung sounds. By designing a 12-layer CNN, spectrogram images are given to the first layer, and the classification process is made. Before the data augment, an average of 60.14% classification performance is obtained. After the data augmentation, an average classification performance of 64.50% is obtained. In addition, in order to compare the performance of the proposed method with other machine learning methods, MFCC features are extracted from the data set samples and the classification process is made with SVM and K-NN.