Signal denoising remains to be one of the main problems in the field of signal processing. Various signal denoising algorithms using wavelet transforms have been introduced. Wavelets show superior signal denoising performance due to their properties such as multiresolution and windowing. This study focuses on denoising of phonocardiogram (PCG) signals using different families of discrete wavelet transforms, thresholding types and techniques, and signal decomposition levels. In particular, we discuss the effect of the chosen wavelet function and wavelet decomposition level on the efficiency of the denoising algorithm. Denoised signals are compared with the original PCG signal to determine the most suitable parameters (wavelet family, level of decomposition, and thresholding type) for the denoising process. The performance of our algorithm is evaluated using the signal-to-noise ratio, percentage root-mean-square difference, and root-mean-square error. The results show that the level of decomposition and thresholding type are the most important parameters affecting the efficiency of the denoising algorithm. Finally, we compare our results with those from other studies to test and optimize the performance of the proposed algorithm.

Denoising of Heart Sound Signals Using Discrete Wavelet Transform

Intelligent computer-aided diagnosis (CAD) systems can enhance the diagnostic capabilities of physicians and reduce the time required for accurate diagnosis. CAD systems could provide physicians with a suggestion about the diagnostic of heart diseases. The objective of this paper is to review the recent published preprocessing, feature extraction and classification techniques and their state of the art of phonocardiogram (PCG) signal analysis. Published literature reviewed in this paper shows the potential of machine learning techniques as a design tool in PCG CAD systems and reveals that the CAD systems for PCG signal analysis are still an open problem. Related studies are compared to their datasets, feature extraction techniques and the classifiers they used. Current achievements and limitations in developing CAD systems for PCG signal analysis using machine learning techniques are presented and discussed. In the light of this review, a number of future research directions for PCG signal analysis are provided.

A review of intelligent systems for heart sound signal analysis.

The paper proposes a new approach to heart activity diagnosis based on Gram polynomials and probabilistic neural networks (PNN). Heart disease recognition is based on the analysis of phonocardiogram (PCG) digital sequences. The PNN provides a powerful tool for proper classification of the input data set. The novelty of the proposed approach lies in a powerful feature extraction based on Gram polynomials and the Fourier transform. The proposed system presents good performance obtaining overall sensitivity of 93%, specificity of 91% and accuracy of 94%, using a public database of over 3000 heart beat sound recordings, classified as normal and abnormal heart sounds. Thus, it can be concluded that Gram polynomials and PNN prove to be a very efficient technique using the PCG signal for characterizing heart diseases.

/pdf/automatic-heart-activity-diagnosis-based-on-gram-polynomials-5427l6nepm.pdf

Automatic heart activity diagnosis based on Gram polynomials and probabilistic neural networks

The auscultation of heart sounds has been for decades a fundamental diagnostic tool in clinical practice. Higher effectiveness can be achieved by recording the corresponding biomedical signal, namely the phonocardiographic signal, and processing it by means of traditional signal processing techniques. An unavoidable processing step is the heart sound segmentation, which is still a challenging task from a technical viewpoint—a limitation of state-of-the-art approaches is the unavailability of trustworthy techniques for the detection of heart sound components. The aim of this work is to design a reliable algorithm for the identification and the classification of heart sounds’ main components. The proposed methodology was tested on a sample population of 24 healthy subjects over 10-min-long simultaneous electrocardiographic and phonocardiographic recordings and it was found capable of correctly detecting and classifying an average of 99.2% of the heart sounds along with their components. Moreover, the delay of each component with respect to the corresponding R-wave peak and the delay among the components of the same heart sound were computed: the resulting experimental values are coherent with what is expected from the literature and what was obtained by other studies.

https://www.mdpi.com/1424-8220/19/8/1868/pdf

A Novel Method for Measuring the Timing of Heart Sound Components through Digital Phonocardiography.

Chronic heart failure represents a global pandemic, currently affecting over 26 million of patients worldwide. It is a major contributor in the death rate of patients with cardiovascular diseases and results in more than 1 million hospitalizations annually in Europe and North America. Methods for chronic heart failure detection can be utilized to act preventive, improve early diagnosis and avoid hospitalizations or even life-threatening situations, thus highly enhance the quality of patient&#x2019;s life. In this paper, we present a machine-learning method for chronic heart failure detection from heart sounds. The method consists of: filtering, segmentation, feature extraction and machine learning. The method was tested with a leave-one-subject-out evaluation technique on data from 122 subjects, gathered in the study. The method achieved 96% accuracy, outperforming a majority classifier for 15 percentage points. More specifically, it detects (recalls) 87% of the chronic heart failure subjects with a precision of 87%. The study confirmed that advanced machine learning applied on real-life sounds recorded with an unobtrusive digital stethoscope can be used for chronic heart failure detection.

Chronic Heart Failure Detection from Heart Sounds Using a Stack of Machine-Learning Classifiers

Comparison of Wavelet Transforms For Denoising And Analysis Of PCG Signal

Heart sounds are multi component non-stationary signals characterized as the normal phonocardiogram (PCG) signals and the pathological PCG signals. PCG is a weak biological signal mixed with strong background noise susceptible to interference from noise. The noise may be added due to various sources. The PCG signal has specific individual characteristics which are considered as a physiological sign in a biometric system. Literatures suggest that the method on time-frequency analysis is known as the trimmed mean spectrogram (TMS). The abnormal murmurs in heart sound can be diagnosed. Another method in time-frequency domain is used in which features are extracted from the TMS containing the distribution of the systolic and diastolic signatures. Probability Neural Networks (PNNs) are used in feature extraction with the acoustic intensities in systole and diastole. These methods can detect accurately the heart disease depending on the applied PCG signal but the result obtained is not optimum. An adaptive neuro-fuzzy inference system (ANFIS) is suggested that can correctly detect the pathological condition of heart.

/pdf/classification-of-pcg-signals-a-survey-2n01f8wdbf.pdf

Classification of PCG Signals: A Survey

The present work proposes a Bio Medical-based system for normal and abnormal heart sound identification based on Discrete Wavelet Transform (DWT), it is useful for Detection of heart diseases. Due to the presence of sampling Frequency components, the wavelets have a different Decomposition level and it's very much important to find, from which level we get the best performance for a particular Heart Sound. DWT is applied up to 10 levels to extract the features for the individual Heart Signal. In this present approach, one dimensional feature extraction is obtained by evaluating parameters like mean energy, variance, standard deviation and maximum entropy etc. Thus we get the proper level of wavelet and using that, we can extract the best feature for Heart Sounds.

Performance analysis of DWT at different levels for feature extraction of PCG signals

Analysis of PCG Signals Using Daubechies Wavelet Family

PCG signals are generated as sound signals by human heart which contains high amount of information related to pathological condition having frequency range of 10 KHz to 250 KHz. Death rates has ranked 2nd all over the world due to undetected pathological condition of heart sound. PCG signals are utilized for the proper recognition of heart disease to save or take proper action based on particular disease. K-Nearest- Neighbor (KNN) is the technique that exploits the concept of clustering the data and identifying the particular class among huge data. The system takes matrix of data about available information on the disease as training data and provides the output as the disease for the sample data provided to the system as input. Many costly methods like echocardiogram already exist for identifying the pathological condition of the heart which is costly and hence the proposed system will provide an efficient tool to get the conditions detected.

Abhishek Misal

Papers

Comparison of Wavelet Transforms For Denoising And Analysis Of PCG Signal

Classification of PCG Signals: A Survey

Performance analysis of DWT at different levels for feature extraction of PCG signals

Analysis of PCG Signals Using Daubechies Wavelet Family

Comparative Evaluation for Classification of PCG Signals using KNN Technique for Wavelet Transforms