Figure - available from: Health Information Science and Systems
This content is subject to copyright. Terms and conditions apply.
Source publication
Background and objective
Heart sound contains various important quantities that help early detection of heart diseases. Many methods have been proposed so far where various signal-processing techniques have been used on heart sounds for heart disease detection.
Methods
In this paper, a methodology is introduced for heart disease detection based on...
Similar publications
The paper aims to study the identification and diagnosis of infrared thermal fault of airborne circuit board of equipment, expand the application of intelligent algorithm in infrared thermal fault diagnosis, and promote the development of computer image processing technology and neural network technology in the field of thermal diagnosis. Taking th...
In this paper, our main aim is to investigate the context-based pixel-wise classification using a fully convolutional neural networks model for flood extent mapping from multispectral remote sensing images. Our approach helps to overcome the limitation of the conventional classification methods with low generalisation ability that used per-pixel sp...
In this research our aim was to identify good borrowers within the context of social lending. We investigated effective ways to conduct this credit risk assessment of these borrowers. We implemented a classification approach in order to make our analysis and came upon the problem of a major class imbalance. In literature, there are few studies on s...
Human physical action classification is an emerging area of research for human-to-machine interaction, which can help to disable people to interact with real world, and robotics application. EMG signals measure the electrical activity muscular systems, which involved in physical action of human. EMG signals provide more information related to physi...
Multi-sensor data fusion is a feasible technique to achieve accurate and robust results in fault diagnosis of rotating machinery under complex conditions. However, the problem of information losses is always ignored during the fusion process. To solve above problem, an ensemble convolutional neural network model is proposed for bearing fault diagno...
Citations
... The model also uses the technique of global average pooling at the end of the network, which helps in the formation of deeper architecture. The average pooling decreases the number of trainable parameters to 0 and leads to an improvement in accuracy [32]. ...
Heart-related diseases that result in heart attacks and cardiac arrhythmias if not detected at an early stage are an increasing concern in the world. Consequently, heart disease detection and classification play an important role in early cardiovascular disease detection to reduce the mortality rate. Abnormal heart patterns like murmurs indicate grave heart conditions and can be detected by trained doctors using a stethoscope. Even though many techniques for heart disease detection based on heart sound processing have been proposed recently, most of the approaches are formulated on conventional features and shallow learning classification algorithms. Machine learning and computational intelligence techniques make a significant contribution to heart disorder disease detection. This chapter presents an improved computer-aided heart disease diagnosis algorithm using phonocardiogram (PCG) signals. The proposed algorithm aims to conduct a performance evaluation of features extracted from three transfer learning models, ResNet-50, GoogleNet, and Inception-V3, trained on the time-frequency visual representation of phonocardiogram heartbeat samples from the cardiology PhysioNet/CinC challenge 2016 database. After the pre-processing of input heart sound, it is converted into spectrogram time-frequency image representation and applied to deep learning models for feature extraction. To identify the most relevant and discard redundant features, the Manta ray foraging optimization algorithm is utilized. The performance of the algorithm proposed for cardiovascular disease detection is evaluated using accuracy, recall, specificity, F1-score, and gmean, and each transfer learning model result is compared using the KNN classifier. It is evident from the experimental results that Manta ray optimized Inception-V3 features provide the highest detection accuracy of 99.58%.
... Figure 24 depicts a clinical PCG signal analysis method a medical practitioner uses. Figure 24 depicts the application of a stethoscope [135,136] in a Heart Sound analysis problem. Heart Sound is acquired from the patient and goes to the preprocessing block for further preprocessing, like filtering, normalization, and segmentation. ...
The paper reviews the milestones and various modern-day approaches in developing phonocardiogram (PCG) signal analysis. It also explains the different phases and methods of the Heart Sound signal analysis. Many physicians depend heavily on ECG experts, inviting healthcare costs and ignorance of stethoscope skills. Hence, auscultation is not a simple solution for the detection of valvular heart disease; therefore, doctors prefer clinical evaluation using Doppler Echo-cardiogram and another pathological test. However, the benefits of auscultation and other clinical evaluation can be associated with computer-aided diagnosis methods that can help considerably in measuring and analyzing various Heart Sounds. This review covers the most recent research for segmenting valvular Heart Sound during preprocessing stages, like adaptive fuzzy system, Shannon energy, time-frequency representation, and discrete wavelet distribution for analyzing and diagnosing various heart-related diseases. Different Convolutional Neural Network (CNN) based deep-learning models are discussed for valvular Heart Sound analysis, like LeNet-5, AlexNet, VGG16, VGG19, DenseNet121, Inception Net, Residual Net, Google Net, Mobile Net, Squeeze Net, and Xception Net. Among all deep-learning methods, the Xception Net claimed the highest accuracy of 99.43 + 0.03% and sensitivity of 98.58 + 0.06%. The review also provides the recent advances in the feature extraction and classification techniques of Cardiac Sound, which helps researchers and readers to a great extent.
... Sometimes pre-trained models are useful for feature extraction. In [24], Demir et al. employed three pre-trained deep CNN models such as AlexNet, VGG16, and VGG19 for feature extraction. Then, the spectrogram is applied to construct input data for the system and finally, SVM is utilized for classification. ...
... In SS-PLSR [18], feature extraction and dimension reduction are achieved through scaledspectrogram and partial least squares regression (PLSR). In Demir et al. [24], different pre-trained models such as AlexNet, VGG16, VGG19 as well as their combinations i.e. AlexNet-VGG16, AlexNet-VGG19, VGG16-VGG19, AlexNet-VGG16-VGG19, etc. are employed for feature extraction. ...
... AlexNet-VGG16, AlexNet-VGG19, VGG16-VGG19, AlexNet-VGG16-VGG19, etc. are employed for feature extraction. SVM-DM [15], SS-PLSR [18], SS-TD [19] and Demir et al. [24] all utilize the Support Vector Machines (SVM) classifier to perform the heart sounds classification. WSCNN [22], CSWT [23] use spectrogram extraction and scaling, Scattering Wavelet Transformation (SWT) for feature representation respectively. ...
Cardiovascular diseases (CVDs) are a group of diseases that affect the heart or blood vessels and are the leading cause of mortality around the world. The main focus of this work is to classify heart sounds accurately before the condition of the heart worsens. Over the past few decades, audio-based classifications of medical data have received high consideration among various researchers. However, there are mainly two drawbacks while working with audio data. (1) variable length of the audio data and, (2) an inadequate number of class quintessential audio samples. In this work, reflection operation and the sliding window approach are employed to generate fixed-sized audio data. The proposed method applied Discrete Wavelet Transform (DWT) and Mel-Spectrogram for feature extraction. Furthermore, different augmentation techniques such as time-stretching and pitch-shifting are utilized in the method so that the proposed deep learning-based CNN architecture can be trained on a large amount of data. The proposed method is verified using two datasets provided by the ‘PASCAL Heart Sounds Challenge’, each of which contains a small number of heart sound samples of various lengths. In comparison, the experimental outcomes exhibit that the proposed approach outperforms many state-of-the-art methods with respect to sensitivity, specificity, precision, Youden index, etc.
... The above-mentioned methods have certain limitations, such as the considerable cost, the time required, and the manual preprocessing of the signals. To overcome these limitations, researchers have used deep learning models like CNNs [24,25] and recurrent neural networks (RNNs) [26,27] to study ECG signals, but their work still relies on the pre-processing of the signals, which can result in information loss. ...
... This did not require prior handcrafted feature extraction. The research in [24,25] demonstrated that DNNs perform better when some temporal variables are used along with raw data inputs. In the literature, many studies have used different datasets: some have used freely accessible ones like the MIT-BIH Arrhythmia Database [17,31] and the PhysioNet Challenge datasets [32], while others have collected and annotated their own data [33]. ...
... A comparative study of the methodologies used in the literature identifies a variety of deep learning methods for the detection of arrhythmias. CNNs have demonstrated remarkable performance in feature extraction from raw ECG signals [24,25]. RNNs excel at modeling temporal relationships, although they can be noise-sensitive and may need a lot of data per-processing step [26,27]. ...
Heart diseases is the world’s principal cause of death, and arrhythmia poses a serious risk to the health of the patient. Electrocardiogram (ECG) signals can be used to detect arrhythmia early and accurately, which is essential for immediate treatment and intervention. Deep learning approaches have played an important role in automatically identifying complicated patterns from ECG data, which can be further used to identify arrhythmia. In this paper, deep-learning-based methods for arrhythmia identification using ECG signals are thoroughly studied and their performances evaluated on the basis of accuracy, specificity, precision, and F1 score. We propose the development of a small CNN, and its performance is compared against pretrained models like GoogLeNet. The comparative study demonstrates the promising potential of deep-learning-based arrhythmia identification using ECG signals.
... Cristhian Potes presented an ensemble of feature-based and deep learning-based classifiers for the detection of abnormal heart sounds [6]. F. Demir designed a method for classifying heart sounds based on convolutional deep neural network (CNN) [7]. T. Li, C. Qing, employed a convolutional neural network (CNN) for classifying heart sounds [8]. ...
This paper introduces an emerging application of artificial intelligence (AI) for the early and remote detection of cardiac murmurs. Using a freely available dataset, the study examines the effectiveness of two most commonly used AI techniques, namely the 2D CNN and LSTM models. The study uses the typical models to compare their relative performance. Subsequently, the better of the two models is utilized and integrated in a web application. The web application enables uploading of the auscultation recordings, it then processes the sound and runs the AI model on it. The output is the classification of sound as normal or with murmur. The experiments for relative comparison of the two AI techniques, demonstrate the superiority of the 2D CNN model over LSTM, as evident through enhanced performance metrics accuracy, recall, and precision. The findings underscore the potential of deep learning algorithms, particularly 2D CNN, in effectively detecting the cardiac murmur detection which can further the cause of improved medical diagnosis and patient care. Keywords - Auscultation, Mel-frequency cepstral coefficient (MFCCs), Convolutional neural network (CNN), Long short term memory (LSTM), cardiac murmurs.
... The classification model was developed using Keras API and Tensor Flow 2.0 in Python 3.7.x. The methodology presented in [37][38][39] inspired this approach. ...
Falls in older people are a major health concern as the leading cause of disability and the second most common cause of accidental death. We developed a rapid fall risk assessment based on a combination of physical performance measurements made with an inertial sensor embedded in a smartphone. This study aimed to evaluate and validate the reliability and accuracy of an easy-to-use smartphone fall risk assessment by comparing it with the Physiological Profile Assessment (PPA) results. Sixty-five participants older than 55 performed a variation of the Timed Up and Go test using smartphone sensors. Balance and gait parameters were calculated, and their reliability was assessed by the (ICC) and compared with the PPAs. Since the PPA allows classification into six levels of fall risk, the data obtained from the smartphone assessment were categorised into six equivalent levels using different parametric and nonparametric classifier models with neural networks. The F1 score and geometric mean of each model were also calculated. All selected parameters showed ICCs around 0.9. The best classifier, in terms of accuracy, was the nonparametric mixed input data model with a 100% success rate in the classification category. In conclusion, fall risk can be reliably assessed using a simple, fast smartphone protocol that allows accurate fall risk classification among older people and can be a useful screening tool in clinical settings.
... The Input 2 the spectrogram image of the signals was used (Ronneberger et al., 2015;Demir et al., 2019), where three 2D convolutional layers with a kernel size of 3 × 3 and ReLu activation functions were concatenated. ...
Introduction: Parkinson's disease is one of the most prevalent neurodegenerative diseases. In the most advanced stages, PD produces motor dysfunction that impairs basic activities of daily living such as balance, gait, sitting, or standing. Early identification allows healthcare personnel to intervene more effectively in rehabilitation. Understanding the altered aspects and impact on the progression of the disease is important for improving the quality of life. This study proposes a two-stage neural network model for the classifying the initial stages of PD using data recorded with smartphone sensors during a modified Timed Up & Go test.
Methods: The proposed model consists on two stages: in the first stage, a semantic segmentation of the raw sensor signals classifies the activities included in the test and obtains biomechanical variables that are considered clinically relevant parameters for functional assessment. The second stage is a neural network with three input branches: one with the biomechanical variables, one with the spectrogram image of the sensor signals, and the third with the raw sensor signals.
Results: This stage employs convolutional layers and long short-term memory. The results show a mean accuracy of 99.64% for the stratified k-fold training/validation process and 100% success rate of participants in the test phase.
Discussion: The proposed model is capable of identifying the three initial stages of Parkinson's disease using a 2-min functional test. The test easy instrumentation requirements and short duration make it feasible for use feasible in the clinical context.
... With the development of computer vision, deep CNN is partially undertaking the analysis task to provide the auxiliary diagnosis. The TFDs were used as inputs to train the deep learning algorithms [21][22][23][24][25][26][27][28][29]. However, due to the different databases, inputs, and network architectures, it is unclear how selecting TFDs, and CNNs can affect heart sound classification. ...
Time-Frequency Distributions (TFDs) support the heart sound characterisation and classification in early cardiac screening. However, despite the frequent use of TFDs in signal analysis, no comprehensive study has been conducted to compare their performances in deep learning for automatic diagnosis. This study is the first to investigate and compare the optimal use of single/combined TFDs for heart sound classification using deep learning. The main contribution of this study is that it provides practical insights into the selection of TFDs as convolutional neural network (CNN) inputs and the design of CNN architecture for heart sound classification. The presented results revealed that: 1) The transformation of the heart sound signal into the TF domain achieves higher classification performance than using raw signal patterns as input. Overall, the difference in the performance was slight among the applied TFDs for all participated CNNs (within 1.3% in MAcc (average of sensitivity and specificity)). However, continuous wavelet transform (CWT) and Chirplet transform (CT) outperformed the rest (surpassing by approximately 0.5−1.3% in MAcc). 2) The appropriate increase of the CNN capacity and architecture optimisation can improve the performance, while the network architecture should not be overly complicated. Based on the results on ResNet or SEResNet, the increasing parameter number and the depth of the structure do not improve the performance apparently. 3) Combining TFDs as CNN inputs did not significantly improve the classification results. The results of this study provide valuable insights for researchers and practitioners in the field of automatic diagnosis of heart sounds with deep learning, particularly in selecting TFDs as CNN input and designing CNN architecture for heart sound classification.
... Although heart sounds are presented as audio signals which are a different data type from ImageNet, DL models trained on ImageNet have shown good performance on time-frequency representations extracted from heart sounds for heart sound classification. Typical DL models on ImageNet, such as AlexNet [135] and VGG [136], have been successfully used for heart sound classification [8], [31], [44], [137]. Compared to ImageNet, AudioSet includes multiple types of acoustic signals and therefore is more close to heart sounds with the consideration of data type. ...
... After extracting representations by pre-trained models, transfer learning uses various types of classifiers for classification, mainly including classic machine learning classifiers and feed-forward neural networks. For instance, SVMs were applied to representations extracted by AlexNet, VGG16, and VGG19 in [31], [44], [137]. Other classifiers such as KNNs were used in [44]. ...
Heart sound auscultation has been demonstrated to be beneficial in clinical usage for early screening of cardiovascular diseases. Due to the high requirement of well-trained professionals for auscultation, automatic auscultation benefiting from signal processing and machine learning can help auxiliary diagnosis and reduce the burdens of training professional clinicians. Nevertheless, classic machine learning is limited to performance improvement in the era of big data. Deep learning has achieved better performance than classic machine learning in many research fields, as it employs more complex model architectures with stronger capability of extracting effective representations. Deep learning has been successfully applied to heart sound analysis in the past years. As most review works about heart sound analysis were given before 2017, the present survey is the first to work on a comprehensive overview to summarise papers on heart sound analysis with deep learning in the past six years 2017--2022. We introduce both classic machine learning and deep learning for comparison, and further offer insights about the advances and future research directions in deep learning for heart sound analysis.
... Compared with traditional ML methods, deep learning algorithms can neglect manual feature extraction and use the raw signal as input with promising performance. After the 2016 PhysioNet/Computing in Cardiology (CinC) Challenge [19], using CNN or RNN to conduct heart sound classification became the mainstream approach [20][21][22][23][24][25]. ...
... With the development of computer vision, deep CNN is partially undertaking the analysis task to provide the auxiliary diagnosis. The TFDs were used as inputs to train the deep learning algorithms [20,[22][23][24][40][41][42][43][44]. However, due to the different databases, inputs, and network architectures, it is unclear how selecting TFDs, and CNNs can affect heart sound classification. ...
