Lung sound classification based on Hilbert-Huang transform features and multilayer perceptron network

Accurate classification of lung sounds plays an important role in noninvasive diagnosis of pulmonary diseases. A novel lung sound classification algorithm based on Hilbert-Huang transform (HHT) features and multilayer perceptron network is proposed in this paper. Three types of HHT domain features, namely the instantaneous envelope amplitude of intrinsic mode functions (IMF), envelop of instantaneous amplitude of the first four layers IMFs, and max value of the marginal spectrum are proposed for jointly characterization of the time-frequency properties of lung sounds. These proposed features are feed into a multi-layer perceptron neural network for training and testing of lung sound signal classification. Abundant experimental work is carried out to verify the effectiveness of the proposed algorithm.