Respiratory sound classification by using an incremental supervised neural network

Determination of lung condition by auscultation is a difficult task and requires special training of medical staff. It is, however, a difficult skill to acquire. In decision making, it is significant to analyze respiratory sounds by an algorithm to give support to medical doctors. In this study, first, a rectangular window is formed so that one cycle of respiratory sound (RS) is contained in this window. Then, the windowed time samples are normalized. In order to extract the features, the normalized RS signal is partitioned into 64 samples of long segments. The power spectrum of each segment is computed, and synchronized summation of power spectra components is performed. Feature vectors are formed by the averaged power spectrum components, yielding 32-dimensional vectors. In the study, classification performances of multi-layer perceptron (MLP), grow and learn (GAL) network and a novel incremental supervised neural network (ISNN) are comparatively examined for the classification of nine different RS classes: Bronchial sounds, bronchovesicular sounds, vesicular sounds, crackles sounds, wheezes sounds, stridor sounds, grunting sounds, squawk sounds, and sounds of friction rub.

[1]  McKUSICK Va,et al.  The acoustic basis of the chest examination; studies by means of sound spectrography. , 1955 .

[2]  Tamer Ölmez,et al.  Classification of Respiratory Sounds by Using an Artificial Neural Network , 2003, Int. J. Pattern Recognit. Artif. Intell..

[3]  Y Homma,et al.  Spectral and waveform characteristics of fine and coarse crackles. , 1991, Thorax.

[4]  Paul H. King,et al.  Representation and Classification of Breath Sounds Recorded in an Intensive Care Setting Using Neural Networks , 2004, Journal of Clinical Monitoring and Computing.

[5]  Robert P. W. Duin,et al.  Pairwise feature evaluation for constructing reduced representations , 2007, Pattern Analysis and Applications.

[6]  Kari Torkkola,et al.  Discriminative features for text document classification , 2003, Formal Pattern Analysis & Applications.

[7]  S. Haltsonen,et al.  Validated method for automatic detection of lung sound crackles , 1991, Medical and Biological Engineering and Computing.

[8]  LinLin Shen,et al.  A review on Gabor wavelets for face recognition , 2006, Pattern Analysis and Applications.

[9]  P Piirilä,et al.  Crackles in patients with fibrosing alveolitis, bronchiectasis, COPD, and heart failure. , 1991, Chest.

[10]  Mineichi Kudo,et al.  Classifier-independent feature selection on the basis of divergence criterion , 2006, Pattern Analysis and Applications.

[11]  E. H. Dooijes,et al.  Classification of Asthmatic Breath Sounds: Preliminary Results of the Classifying Capacity of Human Examiners versus Artificial Neural Networks , 1999, Comput. Biomed. Res..

[12]  J Vanderschoot,et al.  Lung sound intensity in patients with emphysema and in normal subjects at standardised airflows. , 1992, Thorax.

[13]  L Pesu,et al.  Classification of respiratory sounds based on wavelet packet decomposition and learning vector quantization. , 1998, Technology and health care : official journal of the European Society for Engineering and Medicine.

[14]  N. Malmurugan,et al.  Neural classification of lung sounds using wavelet coefficients , 2004, Comput. Biol. Medicine.

[15]  P Piirilä,et al.  Changes in crackle characteristics during the clinical course of pneumonia. , 1992, Chest.

[16]  L.J. Hadjileontiadis,et al.  Separation of discontinuous adventitious sounds from vesicular sounds using a wavelet-based filter , 1997, IEEE Transactions on Biomedical Engineering.

[17]  L Pesu,et al.  Significant differences in flow standardised breath sound spectra in patients with chronic obstructive pulmonary disease, stable asthma, and healthy lungs. , 1995, Thorax.

[18]  Xuelong Li,et al.  General Tensor Discriminant Analysis and Gabor Features for Gait Recognition , 2007, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[19]  A. I. Ethem Alpaydin.  Neural models of incremental supervised and unsupervised learning , 1990 .

[20]  T. D. Bui,et al.  Feature dimensionality reduction for the verification of handwritten numerals , 2004, Pattern Analysis and Applications.

[21]  M. Mussell,et al.  The need for standards in recording and analysing respiratory sounds , 1992, Medical and Biological Engineering and Computing.

[22]  Bülent Sankur,et al.  Two-stage classi cation of respiratory sound patterns , 2003 .

[23]  G. N. Webb,et al.  The acoustic basis of the chest examination; studies by means of sound spectrography. , 1955, American review of tuberculosis.

[24]  Y P Kahya,et al.  Comparison of AR-based algorithms for respiratory sounds classification. , 1994, Computers in biology and medicine.

[25]  Hüseyin Polat,et al.  Combining Neural Network and Genetic Algorithm for Prediction of Lung Sounds , 2005, Journal of Medical Systems.

[26]  Xuelong Li,et al.  Supervised Tensor Learning , 2005, ICDM.