Classification of heart sound signals using a novel deep WaveNet model

Oh, Shu Lih; Vicnesh, Jahmunah; Ooi, Chui Ping; Tan, Rui; Ciaccio, Edward J.; Yamakawa, Toshitaka; Tanabe, Mitsuaki; Kobayashi, Makiko; Acharya, U. Rajendra

doi:10.1016/j.cmpb.2020.105604

Cited by 118 publications

(63 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Merging the DWT and MFCC features, centroid displacement based K-NN, DNN and SVM respectively yields an accuracy of 97.4%, 92.1% and 97.9% [18]. In the same vein, using the same dataset, a recent study has proposed WaveNet, a novel 1D CNN network having residual blocks with dilated 1D CNN and multiple skip connections [19]. Upon 10-fold cross-validation, they have achieved training accuracy of 97% and approximately 90% validation accuracy.…”

Section: A Physiological Origin Of Cardiac Auscultationmentioning

confidence: 99%

“…Although multiple analysis were performed adopting several feature extraction and classifier combinations, MFCC-DWT based conjoined features demonstrated the best performance, giving 97.4%, 97.9% and 92.1% accuracy respectively, on KNN, SVM and DNN classifiers. Another work proposing a novel deep WaveNet architecture has achieved a training accuracy of 97% and validation accuracy of 90% for 5 class CVD classification [19].Furthermore, two recent studies on the classification of HVDs have utilized this dataset dropping the MVP class. The first work has achieved the mean accuracy of 95.13% using wavelet synchrosqueezing transform function (WSST) for extracting the magnitude and phase features from the PCG and classified the signals with the Random Forest (RF) classifier [23] while the second one contains several complex processing steps.…”

Section: Comparison With the Existing Workmentioning

confidence: 99%

“…Among the feature extraction techniques, fast Fourier transform (FFT), short Fourier transform (STFT), discrete wavelet transform (DWT), continuous wavelet transform (CWT), Q wavelet transform (TQWT), chirplet transform (CT), S-transform, Hilbert-Huang transform (HHT), TF decomposition, Empirical mode decomposition (EMD), Mel-Frequency Cepstral Coefficients (MFCCs), Variance Fractal Dimension (VFD) are observed in many works [9]. Despite being quite promising, most of these ML-based algorithm rely on hand-engineered features in the classifier and thus, are not generalized [19]. Advent of deep learning (DL) architectures namely Convolutional Neural Networks (CNNs), Unsupervised Pre-trained networks (UPNs), Recurrent and Recursive neural networks (RNNs) and their application in the the domain of PCG classification have intuitively resolved both the generalization and the accuracy concern by utilizing the inherent self-learning competency of these networks [14], [15].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

CardioXNet: A Novel Lightweight Deep Learning Framework for Cardiovascular Disease Classification Using Heart Sound Recordings

et al. 2021

View full text Add to dashboard Cite

The alarmingly high mortality rate and increasing global prevalence of cardiovascular diseases (CVDs) signify the crucial need for early detection schemes. Phonocardiogram (PCG) signals have been historically applied in this domain owing to its simplicity and cost-effectiveness. In this paper, we propose CardioXNet, a novel lightweight end-to-end CRNN architecture for automatic detection of five classes of cardiac auscultation namely normal, aortic stenosis, mitral stenosis, mitral regurgitation and mitral valve prolapse using raw PCG signal. The process has been automated by the involvement of two learning phases namely, representation learning and sequence residual learning. Three parallel CNN pathways have been implemented in the representation learning phase to learn the coarse and fine-grained features from the PCG and to explore the salient features from variable receptive fields involving 2D-CNN based squeeze-expansion. Thus, in the representation learning phase, the network extracts efficient timeinvariant features and converges with great rapidity. In the sequential residual learning phase, because of the bidirectional-LSTMs and the skip connection, the network can proficiently extract temporal features without performing any feature extraction on the signal. The obtained results demonstrate that the proposed end-toend architecture yields outstanding performance in all the evaluation metrics compared to the previous stateof-the-art methods with up to 99.60% accuracy, 99.56% precision, 99.52% recall and 99.68% F1-score on an average while being computationally comparable. This model outperforms any previous works using the same database by a considerable margin. Moreover, the proposed model was tested on PhysioNet/CinC 2016 challenge dataset achieving an accuracy of 86.57%. Finally the model was evaluated on a merged dataset of Github PCG dataset and PhysioNet dataset achieving excellent accuracy of 88.09%. The high accuracy metrics on both primary and secondary dataset combined with a significantly low number of parameters and end-to-end prediction approach makes the proposed network especially suitable for point of care CVD screening in low resource setups using memory constraint mobile devices.

show abstract

Section: A Physiological Origin Of Cardiac Auscultationmentioning

confidence: 99%

Section: Comparison With the Existing Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

CardioXNet: A Novel Lightweight Deep Learning Framework for Cardiovascular Disease Classification Using Heart Sound Recordings

et al. 2021

View full text Add to dashboard Cite

show abstract

“…The CT-based time-frequency features obtained from PCG and composite classification model yielded an overall accuracy value of 98.33% [ 13 ]. Oh et al [ 79 ] have proposed a waveNet-based DNN model for the classification of HVAs using PCG recordings and obtained an overall accuracy value of 98.20%. The proposed approach demonstrated higher classification performance as compared to the existing algorithms for automated HVA detection.…”

Section: Resultsmentioning

confidence: 99%

Deep Layer Kernel Sparse Representation Network for the Detection of Heart Valve Ailments from the Time-Frequency Representation of PCG Recordings

Ghosh

Ponnalagu

Tripathy

et al. 2020

BioMed Research International

View full text Add to dashboard Cite

The heart valve ailments (HVAs) are due to the defects in the valves of the heart and if untreated may cause heart failure, clots, and even sudden cardiac death. Automated early detection of HVAs is necessary in the hospitals for proper diagnosis of pathological cases, to provide timely treatment, and to reduce the mortality rate. The heart valve abnormalities will alter the heart sound and murmurs which can be faithfully captured by phonocardiogram (PCG) recordings. In this paper, a time-frequency based deep layer kernel sparse representation network (DLKSRN) is proposed for the detection of various HVAs using PCG signals. Spline kernel-based Chirplet transform (SCT) is used to evaluate the time-frequency representation of PCG recording, and the features like L1-norm (LN), sample entropy (SEN), and permutation entropy (PEN) are extracted from the different frequency components of the time-frequency representation of PCG recording. The DLKSRN formulated using the hidden layers of extreme learning machine- (ELM-) autoencoders and kernel sparse representation (KSR) is used for the classification of PCG recordings as normal, and pathology cases such as mitral valve prolapse (MVP), mitral regurgitation (MR), aortic stenosis (AS), and mitral stenosis (MS). The proposed approach has been evaluated using PCG recordings from both public and private databases, and the results demonstrated that an average sensitivity of 100%, 97.51%, 99.00%, 98.72%, and 99.13% are obtained for normal, MVP, MR, AS, and MS cases using the hold-out cross-validation (CV) method. The proposed approach is applicable for the Internet of Things- (IoT-) driven smart healthcare system for the accurate detection of HVAs.

show abstract

“…Table IV shows a comparison of class-wise accuracy between deep wavenet model [24] and the proposed trimodal system. The proposed architecture exhibits better performance in each class.…”

Section: Comparison With Other State Of the Artsmentioning

confidence: 99%

Multimodal CNN Fusion Architecture with Multi-Features for Heart Sound Classification

Ranipa

Zhu

Swamy

2021

2021 IEEE International Symposium on Circuits and Systems (ISCAS)

View full text Add to dashboard Cite

In this paper, a novel multimodal convolutional neural network (CNN) fusion architecture is proposed for heart sound signal classification. Instead of using features from just one domain, general frequency features as well as Mel domain features are extracted from the raw heart sound. The multimodal CNN fusion architecture is individually trained based on the feature maps resulting from various feature extraction methods. These feature maps are then merged for optimizing the diversified extracted features. The proposed method provides an opportunity to explore the optimal selection of features for heart sound classification. Extensive experimentations are carried out, showing that an outstanding accuracy of 98.5% is achieved by the multimodal CNN architecture, which outperforms the other state-of-the-art approaches.

show abstract

Classification of heart sound signals using a novel deep WaveNet model

Cited by 118 publications

References 35 publications

CardioXNet: A Novel Lightweight Deep Learning Framework for Cardiovascular Disease Classification Using Heart Sound Recordings

CardioXNet: A Novel Lightweight Deep Learning Framework for Cardiovascular Disease Classification Using Heart Sound Recordings

Deep Layer Kernel Sparse Representation Network for the Detection of Heart Valve Ailments from the Time-Frequency Representation of PCG Recordings

Multimodal CNN Fusion Architecture with Multi-Features for Heart Sound Classification

Contact Info

Product

Resources

About