Feature-Based Fusion Using CNN for Lung and Heart Sound Classification

Identifying pediatric heart murmurs and distinguishing innocent from pathologic using deep learning

OBJECTIVE

To develop a deep learning algorithm to perform multi-class classification of normal pediatric heart sounds, innocent murmurs, and pathologic murmurs.

METHODS

We prospectively enrolled children under age 18 being evaluated by the Division of Pediatric Cardiology. Parents provided consent for a deidentified recording of their child's heart sounds with a digital stethoscope. Innocent murmurs were validated by a pediatric cardiologist and pathologic murmurs were validated by echocardiogram. To augment our collection of normal heart sounds, we utilized a public database of pediatric heart sound recordings (Oliveira, 2022). We propose two novel approaches for this audio classification task. We train a vision transformer on either Markov transition field or Gramian angular field image representations of the frequency spectrum. We benchmark our results against a ResNet-50 CNN trained on spectrogram images.

RESULTS

Our final dataset consisted of 366 normal heart sounds, 175 innocent murmurs, and 216 pathologic murmurs. Innocent murmurs collected include Still's murmur, venous hum, and flow murmurs. Pathologic murmurs included ventricular septal defect, tetralogy of Fallot, aortic regurgitation, aortic stenosis, pulmonary stenosis, mitral regurgitation and stenosis, and tricuspid regurgitation. We find that the Vision Transformer consistently outperforms the ResNet-50 on all three image representations, and that the Gramian angular field is the superior image representation for pediatric heart sounds. We calculated a one-vs-rest multi-class ROC curve for each of the three classes. Our best model achieves an area under the curve (AUC) value of 0.92 ± 0.05, 0.83 ± 0.04, and 0.88 ± 0.04 for identifying normal heart sounds, innocent murmurs, and pathologic murmurs, respectively.

CONCLUSION

We present two novel methods for pediatric heart sound classification, which outperforms the current standard of using a convolutional neural network trained on spectrogram images. To our knowledge, we are the first to demonstrate multi-class classification of pediatric murmurs. Multiclass output affords a more explainable and interpretable model, which can facilitate further model improvement in the downstream model development cycle and enhance clinician trust and therefore adoption.

Auscultation-Based Pulmonary Disease Detection through Parallel Transformation and Deep Learning

Respiratory diseases are among the leading causes of death, with many individuals in a population frequently affected by various types of pulmonary disorders. Early diagnosis and patient monitoring (traditionally involving lung auscultation) are essential for the effective management of respiratory diseases. However, the interpretation of lung sounds is a subjective and labor-intensive process that demands considerable medical expertise, and there is a good chance of misclassification. To address this problem, we propose a hybrid deep learning technique that incorporates signal processing techniques. Parallel transformation is applied to adventitious respiratory sounds, transforming lung sound signals into two distinct time-frequency scalograms: the continuous wavelet transform and the mel spectrogram. Furthermore, parallel convolutional autoencoders are employed to extract features from scalograms, and the resulting latent space features are fused into a hybrid feature pool. Finally, leveraging a long short-term memory model, a feature from the latent space is used as input for classifying various types of respiratory diseases. Our work is evaluated using the ICBHI-2017 lung sound dataset. The experimental findings indicate that our proposed method achieves promising predictive performance, with average values for accuracy, sensitivity, specificity, and F1-score of 94.16%, 89.56%, 99.10%, and 89.56%, respectively, for eight-class respiratory diseases; 79.61%, 78.55%, 92.49%, and 78.67%, respectively, for four-class diseases; and 85.61%, 83.44%, 83.44%, and 84.21%, respectively, for binary-class (normal vs. abnormal) lung sounds.

Bat2Web: A Framework for Real-Time Classification of Bat Species Echolocation Signals Using Audio Sensor Data

Bats play a pivotal role in maintaining ecological balance, and studying their behaviors offers vital insights into environmental health and aids in conservation efforts. Determining the presence of various bat species in an environment is essential for many bat studies. Specialized audio sensors can be used to record bat echolocation calls that can then be used to identify bat species. However, the complexity of bat calls presents a significant challenge, necessitating expert analysis and extensive time for accurate interpretation. Recent advances in neural networks can help identify bat species automatically from their echolocation calls. Such neural networks can be integrated into a complete end-to-end system that leverages recent internet of things (IoT) technologies with long-range, low-powered communication protocols to implement automated acoustical monitoring. This paper presents the design and implementation of such a system that uses a tiny neural network for interpreting sensor data derived from bat echolocation signals. A highly compact convolutional neural network (CNN) model was developed that demonstrated excellent performance in bat species identification, achieving an F1-score of 0.9578 and an accuracy rate of 97.5%. The neural network was deployed, and its performance was evaluated on various alternative edge devices, including the NVIDIA Jetson Nano and Google Coral.

Dallal Bashi O, Nik Ali N, Mahmood Al Kubaisi Y. Lung disease recognition methods using audio-based analysis with machine learning

The use of computer-based automated approaches and improvements in lung sound recording techniques have made lung sound-based diagnostics even better and devoid of subjectivity errors. Using a computer to evaluate lung sound features more thoroughly with the use of analyzing changes in lung sound behavior, recording measurements, suppressing the presence of noise contaminations, and graphical representations are all made possible by computer-based lung sound analysis. This paper starts with a discussion of the need for this research area, providing an overview of the field and the motivations behind it. Following that, it details the survey methodology used in this work. It presents a discussion on the elements of sound-based lung disease classification using machine learning algorithms. This includes commonly prior considered datasets, feature extraction techniques, pre-processing methods, artifact removal methods, lung-heart sound separation, deep learning algorithms, and wavelet transform of lung audio signals. The study introduces studies that review lung screening including a summary table of these references and discusses the literature gaps in the existing studies. It is concluded that the use of sound-based machine learning in the classification of respiratory diseases has promising results. While we believe this material will prove valuable to physicians and researchers exploring sound-signal-based machine learning, large-scale investigations remain essential to solidify the findings and foster wider adoption within the medical community.

Classification of Adventitious Sounds Combining Cochleogram and Vision Transformers

Early identification of respiratory irregularities is critical for improving lung health and reducing global mortality rates. The analysis of respiratory sounds plays a significant role in characterizing the respiratory system's condition and identifying abnormalities. The main contribution of this study is to investigate the performance when the input data, represented by cochleogram, is used to feed the Vision Transformer (ViT) architecture, since this input-classifier combination is the first time it has been applied to adventitious sound classification to our knowledge. Although ViT has shown promising results in audio classification tasks by applying self-attention to spectrogram patches, we extend this approach by applying the cochleogram, which captures specific spectro-temporal features of adventitious sounds. The proposed methodology is evaluated on the ICBHI dataset. We compare the classification performance of ViT with other state-of-the-art CNN approaches using spectrogram, Mel frequency cepstral coefficients, constant-Q transform, and cochleogram as input data. Our results confirm the superior classification performance combining cochleogram and ViT, highlighting the potential of ViT for reliable respiratory sound classification. This study contributes to the ongoing efforts in developing automatic intelligent techniques with the aim to significantly augment the speed and effectiveness of respiratory disease detection, thereby addressing a critical need in the medical field.

Real-time counting of wheezing events from lung sounds using deep learning algorithms: Implications for disease prediction and early intervention

This pioneering study aims to revolutionize self-symptom management and telemedicine-based remote monitoring through the development of a real-time wheeze counting algorithm. Leveraging a novel approach that includes the detailed labeling of one breathing cycle into three types: break, normal, and wheeze, this study not only identifies abnormal sounds within each breath but also captures comprehensive data on their location, duration, and relationships within entire respiratory cycles, including atypical patterns. This innovative strategy is based on a combination of a one-dimensional convolutional neural network (1D-CNN) and a long short-term memory (LSTM) network model, enabling real-time analysis of respiratory sounds. Notably, it stands out for its capacity to handle continuous data, distinguishing it from conventional lung sound classification algorithms. The study utilizes a substantial dataset consisting of 535 respiration cycles from diverse sources, including the Child Sim Lung Sound Simulator, the EMTprep Open-Source Database, Clinical Patient Records, and the ICBHI 2017 Challenge Database. Achieving a classification accuracy of 90%, the exceptional result metrics encompass the identification of each breath cycle and simultaneous detection of the abnormal sound, enabling the real-time wheeze counting of all respirations. This innovative wheeze counter holds the promise of revolutionizing research on predicting lung diseases based on long-term breathing patterns and offers applicability in clinical and non-clinical settings for on-the-go detection and remote intervention of exacerbated respiratory symptoms.

Interpretation of lung disease classification with light attention connected module

Lung diseases lead to complications from obstructive diseases, and the COVID-19 pandemic has increased lung disease-related deaths. Medical practitioners use stethoscopes to diagnose lung disease. However, an artificial intelligence model capable of objective judgment is required since the experience and diagnosis of respiratory sounds differ. Therefore, in this study, we propose a lung disease classification model that uses an attention module and deep learning. Respiratory sounds were extracted using log-Mel spectrogram MFCC. Normal and five types of adventitious sounds were effectively classified by improving VGGish and adding a light attention connected module to which the efficient channel attention module (ECA-Net) was applied. The performance of the model was evaluated for accuracy, precision, sensitivity, specificity, f1-score, and balanced accuracy, which were 92.56%, 92.81%, 92.22%, 98.50%, 92.29%, and 95.4%, respectively. We confirmed high performance according to the attention effect. The classification causes of lung diseases were analyzed using gradient-weighted class activation mapping (Grad-CAM), and the performances of their models were compared using open lung sounds measured using a Littmann 3200 stethoscope. The experts' opinions were also included. Our results will contribute to the early diagnosis and interpretation of diseases in patients with lung disease by utilizing algorithms in smart medical stethoscopes.

Constrained Synthetic Sampling for Augmentation of Crackle Lung Sounds

Crackles are explosive breathing patterns caused by lung air sacs filling with fluid and act as an indicator for a plethora of pulmonary diseases. Clinical studies suggest a strong correlation between the presence of these adventitious auscultations and mortality rate, especially in pediatric patients, underscoring the importance of their pathological indication. While clinically important, crackles occur rarely in breathing signals relative to other phases and abnormalities of lung sounds, imposing a considerable class imbalance in developing learning methodologies for automated tracking and diagnosis of lung pathologies. The scarcity and clinical relevance of crackle sounds compel a need for exploring data augmentation techniques to enrich the space of crackle signals. Given their unique nature, the current study proposes a crackle-specific constrained synthetic sampling (CSS) augmentation that captures the geometric properties of crackles across different projected object spaces. We also outline a task-agnostic validation methodology that evaluates different augmentation techniques based on their goodness of fit relative to the space of original crackles. This evaluation considers both the separability of the manifold space generated by augmented data samples as well as a statistical distance space of the synthesized data relative to the original. Compared to a range of augmentation techniques, the proposed constrained-synthetic sampling of crackle sounds is shown to generate the most analogous samples relative to original crackle sounds, highlighting the importance of carefully considering the statistical constraints of the class under study.

A Low-Cost AI-Empowered Stethoscope and a Lightweight Model for Detecting Cardiac and Respiratory Diseases from Lung and Heart Auscultation Sounds

Cardiac and respiratory diseases are the primary causes of health problems. If we can automate anomalous heart and lung sound diagnosis, we can improve the early detection of disease and enable the screening of a wider population than possible with manual screening. We propose a lightweight yet powerful model for simultaneous lung and heart sound diagnosis, which is deployable in an embedded low-cost device and is valuable in remote areas or developing countries where Internet access may not be available. We trained and tested the proposed model with the ICBHI and the Yaseen datasets. The experimental results showed that our 11-class prediction model could achieve 99.94% accuracy, 99.84% precision, 99.89% specificity, 99.66% sensitivity, and 99.72% F1 score. We designed a digital stethoscope (around USD 5) and connected it to a low-cost, single-board-computer Raspberry Pi Zero 2W (around USD 20), on which our pretrained model can be smoothly run. This AI-empowered digital stethoscope is beneficial for anyone in the medical field, as it can automatically provide diagnostic results and produce digital audio records for further analysis.

A Computer-Aided Heart Valve Disease Diagnosis System Based on Machine Learning

Cardiac auscultation is a noninvasive, convenient, and low-cost diagnostic method for heart valvular disease, and it can diagnose the abnormality of the heart valve at an early stage. However, the accuracy of auscultation relies on the professionalism of cardiologists. Doctors in remote areas may lack the experience to diagnose correctly. Therefore, it is necessary to design a system to assist with the diagnosis. This study proposed a computer-aided heart valve disease diagnosis system, including a heart sound acquisition module, a trained model for diagnosis, and software, which can diagnose four kinds of heart valve diseases. In this study, a training dataset containing five categories of heart sounds was collected, including normal, mitral stenosis, mitral regurgitation, and aortic stenosis heart sound. A convolutional neural network GoogLeNet and weighted KNN are used to train the models separately. For the model trained by the convolutional neural network, time series heart sound signals are converted into time-frequency scalograms based on continuous wavelet transform to adapt to the architecture of GoogLeNet. For the model trained by weighted KNN, features from the time domain and time-frequency domain are extracted manually. Then feature selection based on the chi-square test is performed to get a better group of features. Moreover, we designed software that lets doctors upload heart sounds, visualize the heart sound waveform, and use the model to get the diagnosis. Model assessments using accuracy, sensitivity, specificity, and F1 score indicators are done on two trained models. The results showed that the model trained by modified GoogLeNet outperformed others, with an overall accuracy of 97.5%. The average accuracy, sensitivity, specificity, and F1 score for diagnosing four kinds of heart valve diseases are 98.75%, 96.88%, 99.22%, and 97.99%, respectively. The computer-aided diagnosis system, with a heart sound acquisition module, a diagnostic model, and software, can visualize the heart sound waveform and show the reference diagnostic results. This can assist in the diagnosis of heart valve diseases, especially in remote areas, which lack skilled doctors.

An accurate deep learning model for wheezing in children using real world data

Auscultation is an important diagnostic method for lung diseases. However, it is a subjective modality and requires a high degree of expertise. To overcome this constraint, artificial intelligence models are being developed. However, these models require performance improvements and do not reflect the actual clinical situation. We aimed to develop an improved deep-learning model learning to detect wheezing in children, based on data from real clinical practice. In this prospective study, pediatric pulmonologists recorded and verified respiratory sounds in 76 pediatric patients who visited a university hospital in South Korea. In addition, structured data, such as sex, age, and auscultation location, were collected. Using our dataset, we implemented an optimal model by transforming it based on the convolutional neural network model. Finally, we proposed a model using a 34-layer residual network with the convolutional block attention module for audio data and multilayer perceptron layers for tabular data. The proposed model had an accuracy of 91.2%, area under the curve of 89.1%, precision of 94.4%, recall of 81%, and F1-score of 87.2%. The deep-learning model proposed had a high accuracy for detecting wheeze sounds. This high-performance model will be helpful for the accurate diagnosis of respiratory diseases in actual clinical practice.

Special Issue "Computer Aided Diagnosis Sensors"

Sensors used to diagnose, monitor or treat diseases in the medical domain are known as medical sensors [...].

Diagnosis of cardiac abnormalities based on phonocardiogram using a novel fuzzy matching feature extraction method

BACKGROUND

The diagnosis of cardiac abnormalities based on heart sound signal is a research hotspot in recent years. The early diagnosis of cardiac abnormalities has a crucial significance for the treatment of heart diseases.

METHODS

For the sake of achieving more practical clinical applications of automatic recognition of cardiac abnormalities, here we proposed a novel fuzzy matching feature extraction method. First of all, a group of Gaussian wavelets are selected and then optimized based on a template signal. Convolutional features of test signal and the template signal are then computed. Matching degree and matching energy features between template signal and test signal in time domain and frequency domain are then extracted. To test performance of proposed feature extraction method, machine learning algorithms such as K-nearest neighbor, support vector machine, random forest and multilayer perceptron with grid search parameter optimization are constructed to recognize heart disease using the extracted features based on phonocardiogram signals.

RESULTS

As a result, we found that the best classification accuracy of random forest reaches 96.5% under tenfold cross validation using the features extracted by the proposed method. Further, Mel-Frequency Cepstral Coefficients of phonocardiogram signals combing with features extracted by our algorithm are evaluated. Accuracy, sensitivity and specificity of integrated features reaches 99.0%, 99.4% and 99.7% respectively when using support vector machine, which achieves the best performance among all reported algorithms based on the same dataset. On several common features, we used independent sample t-tests. The results revealed that there are significant differences (p < 0.05) between 5 categories.

CONCLUSION

It can be concluded that our proposed fuzzy matching feature extraction method is a practical approach to extract powerful and interpretable features from one-dimensional signals for heart sound diagnostics and other pattern recognition task.

Design of Abnormal Heart Sound Recognition System Based on HSMM and Deep Neural Network

Introduction

Heart sound signal is an important physiological signal of human body, and the identification and research of heart sound signal is of great significance.

Methods

For abnormal heart sound signal recognition, an abnormal heart sound recognition system, combining hidden semi-Markov models (HSMM) with deep neural networks, is proposed. Firstly, HSMM is used to build a heart sound segmentation model to accurately segment the heart sound signal, and then the segmented heart sound signal is subjected to feature extraction. Finally, the trained deep neural network model is used for recognition.

Results

Compared with other methods, this method has a relatively small amount of input feature data and high accuracy, fast recognition speed.

Discussion

HSMM combined with deep neural network is expected to be deployed on smart mobile devices for telemedicine detection.