An intelligent pattern recognition system based on neural network and wavelet decomposition for interpretation of heart sounds

Abstract

Abstract-In this study, we develop a new automated pattern recognition system for interpretation of heart sound based on wavelet decomposition of signals and classification using neural network. Inputs to the system are the heart sound signals acquired by a stethoscope in a noiseless environment. We generate features for the objective concise representation of heart sound signals by means of wavelet decomposition. Classification of the features is performed using a back propagation neural network with adaptive learning rate. With two hundred record windows obtained from young humans are studied. One hundred of the record windows in database are selected for use as training phase for neural network. In the test result of the intelligent pattern recognition system with ten different types heart sound signals are acquired a high success.

Applying Advanced NN-based Decision Support Scheme for Heart Diseases Diagnosis

“…Overall, the results revealed that all heart diseases of MS, AS and VSD are successfully and accurately identified with the proposed system, as shown in (Fig.1, Fig.2, Fig.3, Fig.4, Fig.5, Fig.6). Compared with previous studies presented in [9,13,15,18] which are mainly focused on valvular heart diseases, the proposed system delivers higher classification accuracy for both valvular and non-valvular heart diseases.…”

“…Turkoglu and Arslan present "An Intelligent Pattern Recognition System Based on Neural Network and Wavelet Decomposition for interpretation of heart sounds" study. The study deals with aortic stenosis, mitral regurgitation, atrial septal defect mitral stenosis and aortic regurgitation disease diagnosis [15].…”

A Survey of Audio Processing Algorithms for Digital Stethoscopes

“…The recognition rate was 95%. Turkoglu (Turkoglu and Arslan, 2001), Ozgur (Ozgur Say and Olmez, 2002) and also used wavelets as feature vectors for classification, although they provide too few details regarding the used data sets. Trimmed mean spectrograms are used by Leung (T.S.…”

Classification of Four Types of Common Murmurs using Wavelets and a Learning Vector Quantization Network

“…The mother wavelet used was Daubechies. Past research have indicated that level 4 and 5 approximations are able to capture important aspects of the murmurs under study, namely aortic and mitral stenosis and aortic and mitral insufficiency [11], [12]. Note also in Figure 4, that the noise level within the murmurs, S1 (systolic) and S2 (diastolic) components is reduced significantly due to filtering.…”

“…The reasons partly have been the lack of further pursuit, unproven or low accuracy rates and primarily unrealizability in the hardware level because of the size of the algorithm. Some design methods use segmentation, wavelets, neural networks, S-transform, Hilbert Transform, decision tree, Hidden Markov models, etc [5][6][7][8][9][10][11][12][13][14][15][16][17]. Despite having designed a variety of algorithms, researchers faced a challenging situation in terms of classification of the heart sounds because of the large number of possible cases as well as the characteristics of the sound from a signal processing perspective.…”