Hidden Markov model with heart sound signals for identification of heart diseases

Heart auscultation still remains a dominant method for the diagnosis of heart diseases caused by heart valve abnormalities. But it is very subjective and significantly relies on the interpretation or perception of well-trained physicians. Thus it would be very desirable to develop a computer-aided automated or semi-automated heart sound identification system that can provide more objective diagnositic results. Recently a hidden Markov model (HMM) has been used quite successfully for the classification of heart sounds. In this paper, we have investigated the classification performance of the MFCC-based HMM with heart sound signals by varying the model’s number of states, number of mixtures, and analysis frame size in MFCC feature extraction. We carried out the classification experiments using the 325 heart sound data made up of 10 different types of heart sound signals. From this, maximum correct classification rate of 95.08% was achieved when the HMM has 4 states, 8 mixtures with analsys frame size of 20ms for feature extraction.

[1]  J. Hertzberg,et al.  Artificial Neural Network-Based Method of Screening Heart Murmurs in Children , 2001, Circulation.

[2]  I. Cathers,et al.  Neural network assisted cardiac auscultation , 1995, Artif. Intell. Medicine.

[3]  N. Hamdy,et al.  Heart diseases diagnosis using HMM , 2002, 11th IEEE Mediterranean Electrotechnical Conference (IEEE Cat. No.02CH37379).

[4]  Paul Mark Bentley,et al.  Time-frequency analysis of native and prosthetic heart valve sounds , 1996 .

[5]  R. Guardo,et al.  Time-frequency analysis of heart murmurs. Part II: Optimisation of time-frequency representations and performance evaluation , 2006, Medical and Biological Engineering and Computing.

[6]  Ping Wang,et al.  A computer-aided MFCC-based HMM system for automatic auscultation , 2008, Comput. Biol. Medicine.

[7]  Curt DeGroff,et al.  A classifier based on the artificial neural network approach for cardiologic auscultation in pediatrics , 2005, Artif. Intell. Medicine.

[8]  Paul R. White,et al.  Classification of heart sounds using time-frequency method and artificial neural networks , 2000, Proceedings of the 22nd Annual International Conference of the IEEE Engineering in Medicine and Biology Society (Cat. No.00CH37143).

[9]  Friedberg Ck.  Computers in cardiology. , 1970 .

[10]  Anthony D. Ricke,et al.  Automatic segmentation of heart sound signals using hidden markov models , 2005, Computers in Cardiology, 2005.

[11]  Aubrey Leatham,et al.  Auscultation of the Heart and Phonocardiography , 1971, Journal of the Royal College of Physicians of London.

[12]  E. Toft,et al.  Segmentation of heart sound recordings from an electronic stethoscope by a duration dependent Hidden-Markov Model , 2008, 2008 Computers in Cardiology.

[13]  G. Castellanos-Dominguez,et al.  Selection of Dynamic Features Based on Time–Frequency Representations for Heart Murmur Detection from Phonocardiographic Signals , 2009, Annals of Biomedical Engineering.

[14]  Yongjoo Chung,et al.  A Classification Approach for the Heart Sound Signals Using Hidden Markov Models , 2006, SSPR/SPR.