ConvLSNet: A lightweight architecture based on ConvLSTM model for the classification of pulmonary conditions using multichannel lung sound recordings.

Characterization of lung sounds (LS) is indispensable for diagnosing respiratory pathology. Although conventional neural networks (NNs) have been widely employed for the automatic diagnosis of lung sounds, deep neural networks can potentially be more useful than conventional NNs by allowing accurate classification without requiring preprocessing and feature extraction. Utilizing the long short-term memory (LSTM) layers to reveal the sequence-based properties of the LS time series, a novel architecture consisting of a cascade of convolutional long short-term memory (ConvLSTM) and LSTM layers, namely ConvLSNet is developed, which permits highly accurate diagnosis of pulmonary disease states. By modeling the multichannel lung sounds through the ConvLSTM layer, the proposed ConvLSNet architecture can concurrently deal with the spatial and temporal properties of the six-channel LS recordings without heavy preprocessing or data transformation. Notably, the proposed model achieves a classification accuracy of 97.4 % based on LS data corresponding to three pulmonary conditions, namely asthma, COPD, and the healthy state. Compared with architectures consisting exclusively of CNN or LSTM layers, as well as those employing a cascade integration of 2DCNN and LSTM layers, the proposed ConvLSNet architecture exhibited the highest classification accuracy, while imposing the lowest computational cost as quantified by the number of parameters, training time, and learning rate.

Tunable Q-factor wavelet transform based lung signal decomposition and statistical feature extraction for effective lung disease classification.

The auscultation is a non-invasive and cost-effective method used for the diagnosis of lung diseases, which are one of the leading causes of death worldwide. However, the efficacy of the auscultation suffers from the limitations of the analog stethoscopes and the subjective nature of human interpretation. To overcome these limitations, the accurate diagnosis of these diseases by employing the computer based automated algorithms applied to the digitized lung sounds has been studied for the last decades. This study proposes a novel approach that uses a Tunable Q-factor Wavelet Transform (TQWT) based statistical feature extraction followed by individual and ensemble learning model training with the aim of lung disease classification. During the learning stage various machine learning algorithms are utilized as the individual learners as well as the hard and soft voting fusion approaches are employed for performance enhancement with the aid of the predictions of individual models. For an objective evaluation of the proposed approach, the study was structured into two main tasks that were investigated in detail by using several sub-tasks to comparison with state-of-the-art studies. Among the sub-tasks which investigates patient-based classification, the highest accuracy obtained for the binary classification was achieved as 97.63% (healthy vs. non-healthy), while accuracy values up to 66.32% for three-class classification (obstructive-related, restrictive-related, and healthy), and 53.42% for five-class classification (asthma, chronic obstructive pulmonary disease, interstitial lung disease, pulmonary infection, and healthy) were obtained. Regarding the other sub-task, which investigates sample-based classification, the proposed approach was superior to almost all previous findings. The proposed method underscores the potential of TQWT based signal decomposition that leverages the power of its adaptive time-frequency resolution property satisfied by Q-factor adjustability. The obtained results are very promising and the proposed approach paves the way for more accurate and automated digital auscultation techniques in clinical settings.

An open auscultation dataset for machine learning-based respiratory diagnosis studies.

Machine learning enabled auscultating diagnosis can provide promising solutions especially for prescreening purposes. The bottleneck for its potential success is that high-quality datasets for training are still scarce. An open auscultation dataset that consists of samples and annotations from patients and healthy individuals is established in this work for the respiratory diagnosis studies with machine learning, which is of both scientific importance and practical potential. A machine learning approach is examined to showcase the use of this new dataset for lung sound classifications with different diseases. The open dataset is available to the public online.

Respiratory sound classification for crackles, wheezes, and rhonchi in the clinical field using deep learning.

Auscultation has been essential part of the physical examination; this is non-invasive, real-time, and very informative. Detection of abnormal respiratory sounds with a stethoscope is important in diagnosing respiratory diseases and providing first aid. However, accurate interpretation of respiratory sounds requires clinician's considerable expertise, so trainees such as interns and residents sometimes misidentify respiratory sounds. To overcome such limitations, we tried to develop an automated classification of breath sounds. We utilized deep learning convolutional neural network (CNN) to categorize 1918 respiratory sounds (normal, crackles, wheezes, rhonchi) recorded in the clinical setting. We developed the predictive model for respiratory sound classification combining pretrained image feature extractor of series, respiratory sound, and CNN classifier. It detected abnormal sounds with an accuracy of 86.5% and the area under the ROC curve (AUC) of 0.93. It further classified abnormal lung sounds into crackles, wheezes, or rhonchi with an overall accuracy of 85.7% and a mean AUC of 0.92. On the other hand, as a result of respiratory sound classification by different groups showed varying degree in terms of accuracy; the overall accuracies were 60.3% for medical students, 53.4% for interns, 68.8% for residents, and 80.1% for fellows. Our deep learning-based classification would be able to complement the inaccuracies of clinicians' auscultation, and it may aid in the rapid diagnosis and appropriate treatment of respiratory diseases.

Childhood wheezing phenotypes and their association to perinatal factors: a longitudinal study.

INTRODUCTION: Multiple gestational and early life factors have been described as the variables that increase the risk for each phenotype of infantile wheezing. Our objective was to study the evolution of wheezing in a cohort of children followed up to 9-10 years of age and its relationship with different perinatal risk factors. METHODS: A longitudinal study was made on the evolution of wheezing, over time, in 1164 children from Salamanca (Spain) included in the International Study of Wheezing in Infants, when the children were 12 months old. They were classified into three phenotypes: transient early wheezing (last episode before 3 years of age), early persistent wheezing (start before 3 years age and persisting thereafter), and late-onset wheezing (first episode after 3 years of age). Univariate and multivariable analyses were performed to establish associations between the different phenotypes and perinatal factors. RESULTS: Data were obtained corresponding to a total of 531 children. Of these, 169 (31.8%) had experienced transient early wheezing, 100 (18.8%) early persistent wheezing, 28 (5.3%) late-onset wheezing, and 234 (44.1%) had never experienced wheezing. Cesarean delivery, early exposure to infections, the presence of atopic eczema, and a smoking father were associated with transient early wheezing. Early persistent wheezing was associated with a family history of allergy, smoking, and obstetric diseases. Exclusive breastfeeding was identified as a protective factor in both transient and persistent early wheezing. Late-onset wheezing was associated with the male gender and with maternal history of rhinitis and eczema. CONCLUSIONS: Wheezing phenotypes were associated with different risk perinatal factors. Knowledge in the field is essential in order to influence the modifiable factors.

Recurrent Severe Preschool Wheeze: From Prespecified Diagnostic Labels to Underlying Endotypes.

Rationale: Preschool wheezing is heterogeneous, but the underlying mechanisms are poorly understood.Objectives: To investigate lower airway inflammation and infection in preschool children with different clinical diagnoses undergoing elective bronchoscopy and BAL.Methods: We recruited 136 children aged 1-5 years (105 with recurrent severe wheeze [RSW]; 31 with nonwheezing respiratory disease [NWRD]). Children with RSW were assigned as having episodic viral wheeze (EVW) or multiple-trigger wheeze (MTW). We compared lower airway inflammation and infection in different clinical diagnoses and undertook data-driven analyses to determine clusters of pathophysiological features, and we investigated their relationships with prespecified diagnostic labels.Measurements and Main Results: Blood eosinophil counts and percentages and allergic sensitization were significantly higher in children with RSW than in children with a NWRD. Blood neutrophil counts and percentages, BAL eosinophil and neutrophil percentages, and positive bacterial culture and virus detection rates were similar between groups. However, pathogen distribution differed significantly, with higher detection of rhinovirus in children with RSW and higher detection of Moraxella in sensitized children with RSW. Children with EVW and children with MTW did not differ in terms of blood or BAL-sample inflammation, or bacteria or virus detection. The Partition around Medoids algorithm revealed four clusters of pathophysiological features: 1) atopic (17.9%), 2) nonatopic with a low infection rate and high use of inhaled corticosteroids (31.3%), 3) nonatopic with a high infection rate (23.1%), and 4) nonatopic with a low infection rate and no use of inhaled corticosteroids (27.6%). Cluster allocation differed significantly between the RSW and NWRD groups (RSW was evenly distributed across clusters, and 60% of the NWRD group was assigned to cluster 4; P < 0.001). There was no difference in cluster membership between the EVW and MTW groups. Cluster 1 was dominated by Moraxella detection (P = 0.04), and cluster 3 was dominated by Haemophilus or Staphylococcus or Streptococcus detection (P = 0.02).Conclusions: We identified four clusters of severe preschool wheeze, which were distinguished by using sensitization, peripheral eosinophilia, lower airway neutrophilia, and bacteriology.

Methods for Adventitious Respiratory Sound Analyzing Applications Based on Smartphones: A Survey.

Detection and classification of adventitious acoustic lung sounds plays an important role in diagnosing, monitoring, controlling and, caring the patients with lung diseases. Such systems can be presented as different platforms like medical devices, standalone software or smartphone application. Ubiquity of smartphones and widespread use of the corresponding applications make such a device an attractive platform for hosting the detection and classification systems for adventitious lung sounds. In this paper, the smartphone-based systems for automatic detection and classification of the adventitious lung sounds are surveyed. Such adventitious sounds include cough, wheeze, crackle and, snore. Relevant sounds related to abnormal respiratory activities are considered as well. The methods are shortly described and the analyzing algorithms are explained. The analysis includes detection and/or classification of the sound events. A summary of the main surveyed methods together with the classification parameters and used features for the sake of comparison is given. Existing challenges, open issues and future trends will be discussed as well.

La auscultación pulmonar en el siglo 21 / Lung auscultation in the 21th century

La auscultación pulmonar es parte fundamental del examen físico para el diagnóstico de las enfermedades respiratorias. La estandarización que ha alcanzado la nomenclatura de los ruidos respiratorios, sumado a los avances en el análisis computacional de los mismos, han permitido mejorar la utilidad de esta técnica. Sin embargo, el rendimiento de la auscultación pulmonar ha sido cuestionado por tener una concordancia variable entre profesionales de la salud. Aun cuando la incorporación de nuevas herramientas diagnósticas de imágenes y función pulmonar han revolucionado la precisión diagnóstica en enfermedades respiratorias, no existe tecnología que permita reemplazar la técnica de auscultación pulmonar para guiar el proceso diagnóstico. Por una parte, la auscultación pulmonar permite seleccionar a aquellos pacientes que se beneficiarán de una determinada técnica diagnóstica, se puede repetir cuantas veces sea necesario para tomar decisiones clínicas, y frecuentemente permite prescindir de exámenes adicionales que no siempre son fáciles de realizar o no se encuentran disponibles. En esta revisión se presenta el estado actual de la técnica de auscultación pulmonar y su rendimiento objetivo basado en la nomenclatura actual aceptada para los ruidos respiratorios, además de resumir la evidencia principal de estudios de concordancia de auscultación pediátrica y su análisis objetivo a través de nueva tecnología computacional.

Lung auscultation is an essential part of the physical examination for diagnosing respiratory diseases. The terminology standardization for lung sounds, in addition to advances in their analysis through new technologies, have improved the use of this technique. However, traditional auscultation has been questioned due to the limited concordance among health professionals. Despite the revolu tionary use of new diagnostic tools of imaging and lung function tests allowing diagnostic accuracy in respiratory diseases, no technology can replace lung auscultation to guide the diagnostic process. Lung auscultation allows identifying those patients who may benefit from a specific test. Moreover, this technique can be performed many times to make clinical decisions, and often with no need for- complicated and sometimes unavailable tests. This review describes the current state-of-the-art of lung auscultation and its efficacy based on the current respiratory sound terminology. In addition, it describes the main evidence on respiratory sound concordance studies among health professionals and its objective analysis through new technology.

Adaptive Filtering Improved Apnea Detection Performance Using Tracheal Sounds in Noisy Environment: A Simulation Study.

OBJECTIVE: Tracheal sounds were used to detect apnea on various occasions. However, ambient noises can contaminate tracheal sounds which result in poor performance of apnea detection. The objective of this paper was to apply the adaptive filtering (AF) algorithm to improve the quality of tracheal sounds and examine the accuracy of the apnea detection algorithm using tracheal sounds after AF. METHOD: Tracheal sounds were acquired using a primary microphone encased in a plastic bell, and the ambient noises were collected using a reference microphone resting outside the plastic bell in quiet and noisy environments, respectively. Simultaneously, the flow pressure signals and thoracic and abdominal movement were obtained as the standard signals to determine apnea events. Then, the normalized least mean square (NLMS) AF algorithm was applied to the tracheal sounds mixed with noises. Finally, the algorithm of apnea detection was used to the tracheal sounds with AF and the tracheal sounds without AF. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), accuracy, and Cohen's kappa coefficient of apnea detection were calculated. RESULTS: Forty-six healthy subjects, aged 18-35 years and with BMI < 21.4, were included in the study. The apnea detection performance using tracheal sounds was as follows: in the quiet environment, the tracheal sounds without AF detected apnea with 97.2% sensitivity, 99.9% specificity, 99.8% PPV, 99.4% NPV, 99.5% accuracy, and 0.982 kappa coefficient. The tracheal sounds with AF detected apnea with 98.2% sensitivity, 99.9% specificity, 99.4% PPV, 99.6% NPV, 99.6% accuracy, and 0.985 kappa coefficient. While in the noisy environment, the tracheal sounds without AF detected apnea with 81.1% sensitivity, 96.9% specificity, 85.1% PPV, 96% NPV, 94.2% accuracy, and 0.795 kappa coefficient and the tracheal sounds with AF detected apnea with 91.5% sensitivity, 97.4% specificity, 88.4% PPV, 98.2% NPV, 96.4% accuracy, and 0.877 kappa coefficient. CONCLUSION: The performance of apnea detection using tracheal sounds with the NLMS AF algorithm in the noisy environment proved to be accurate and reliable. The AF technology could be applied to the respiratory monitoring using tracheal sounds.

Deep Neural Network for Respiratory Sound Classification in Wearable Devices Enabled by Patient Specific Model Tuning.

The primary objective of this paper is to build classification models and strategies to identify breathing sound anomalies (wheeze, crackle) for automated diagnosis of respiratory and pulmonary diseases. In this work we propose a deep CNN-RNN model that classifies respiratory sounds based on Mel-spectrograms. We also implement a patient specific model tuning strategy that first screens respiratory patients and then builds patient specific classification models using limited patient data for reliable anomaly detection. Moreover, we devise a local log quantization strategy for model weights to reduce the memory footprint for deployment in memory constrained systems such as wearable devices. The proposed hybrid CNN-RNN model achieves a score of [Formula: see text] on four-class classification of breathing cycles for ICBHI'17 scientific challenge respiratory sound database. When the model is re-trained with patient specific data, it produces a score of [Formula: see text] for leave-one-out validation. The proposed weight quantization technique achieves ≈ 4 × reduction in total memory cost without loss of performance. The main contribution of the paper is as follows: Firstly, the proposed model is able to achieve state of the art score on the ICBHI'17 dataset. Secondly, deep learning models are shown to successfully learn domain specific knowledge when pre-trained with breathing data and produce significantly superior performance compared to generalized models. Finally, local log quantization of trained weights is shown to be able to reduce the memory requirement significantly. This type of patient-specific re-training strategy can be very useful in developing reliable long-term automated patient monitoring systems particularly in wearable healthcare solutions.

[Lung auscultation in the 21th century]. / La auscultación pulmonar en el siglo 21.

Lung auscultation is an essential part of the physical examination for diagnosing respiratory diseases. The terminology standardization for lung sounds, in addition to advances in their analysis through new technologies, have improved the use of this technique. However, traditional auscultation has been questioned due to the limited concordance among health professionals. Despite the revolu tionary use of new diagnostic tools of imaging and lung function tests allowing diagnostic accuracy in respiratory diseases, no technology can replace lung auscultation to guide the diagnostic process. Lung auscultation allows identifying those patients who may benefit from a specific test. Moreover, this technique can be performed many times to make clinical decisions, and often with no need for- complicated and sometimes unavailable tests. This review describes the current state-of-the-art of lung auscultation and its efficacy based on the current respiratory sound terminology. In addition, it describes the main evidence on respiratory sound concordance studies among health professionals and its objective analysis through new technology.

Blind Equalization of Lung Crackle Sounds to Compensate Chest Attenuation.

Diseased lungs generate adventitious sounds that propagate through the thorax, reaching the surface where they may be heard or recorded. The attenuation imposed to the lung sounds by the thorax depends on the physical characteristics of each patient, hampering the analysis of quantitative indexes measured to assist the diagnosis of cardiorespiratory disorders. This work proposes the application of a blind equalizer (eigenvector algorithm - EVA) to reduce the effects of thorax attenuation on indexes measured from crackle sounds. Computer simulated crackles (acquired on the posterior chest wall after being applied to volunteer's mouth) and actual crackles belonging to a database were equalized. Quantitative indexes were measured from crackles before and after equalization. Comparison of indexes measured from simulated crackles reveals that the equalizer improves the results due to attenuation compensation and removal of Gaussian noise. Effects of equalization on indexes measured from actual crackles were qualitatively assessed. Results point out that blind equalization of crackles recorded on the thorax provides more consistent quantitative indexes to assist the diagnosis of different cardiorespiratory diseases.

Evaluation of features for classification of wheezes and normal respiratory sounds.

Chronic Respiratory Diseases (CRDs), such as Asthma and Chronic Obstructive Pulmonary Disease (COPD), are leading causes of deaths worldwide. Although both Asthma and COPD are not curable, they can be managed by close monitoring of symptoms to prevent worsening of the condition. One key symptom that needs to be monitored is the occurrence of wheezing sounds during breathing since its early identification could prevent serious exacerbations. Since wheezing can happen randomly without warning, a long-term monitoring system with automatic wheeze detection could be extremely helpful to manage these respiratory diseases. This study evaluates the discriminatory ability of different types of feature used in previous related studies, with a total size of 105 individual features, for automatic identification of wheezing sound during breathing. A linear classifier is used to determine the best features for classification by evaluating several performance metrics, including ranksum statistical test, area under the sensitivity--specificity curve (AUC), F1 score, Matthews Correlation Coefficient (MCC), and relative computation time. Tonality index attained the highest effect size, at 87.95%, and was found to be the feature with the lowest p-value when ranksum significance test was performed. Third MFCC coefficient achieved the highest AUC and average optimum F1 score at 0.8919 and 82.67% respectively, while the highest average optimum MCC was obtained by the first coefficient of a 6th order LPC. The best possible combination of two and three features for wheeze detection is also studied. The study concludes with an analysis of the different trade-offs between accuracy, reliability, and computation requirements of the different features since these will be highly useful for researchers when designing algorithms for automatic wheeze identification.

Distinguishing Wheezing Phenotypes from Infancy to Adolescence. A Pooled Analysis of Five Birth Cohorts.

Rationale: Pooling data from multiple cohorts and extending the time frame across childhood should minimize study-specific effects, enabling better characterization of childhood wheezing. Objectives: To analyze wheezing patterns from early childhood to adolescence using combined data from five birth cohorts. Methods: We used latent class analysis to derive wheeze phenotypes among 7,719 participants from five birth cohorts with complete report of wheeze at five time periods. We tested the associations of derived phenotypes with late asthma outcomes and lung function, and investigated the uncertainty in phenotype assignment. Results: We identified five phenotypes: never/infrequent wheeze (52.1%), early onset preschool remitting (23.9%), early onset midchildhood remitting (9%), persistent (7.9%), and late-onset wheeze (7.1%). Compared with the never/infrequent wheeze, all phenotypes had higher odds of asthma and lower forced expiratory volume in 1 second and forced expiratory volume in 1 second/forced vital capacity in adolescence. The association with asthma was strongest for persistent wheeze (adjusted odds ratio, 56.54; 95% confidence interval, 43.75-73.06). We observed considerable within-class heterogeneity at the individual level, with 913 (12%) children having low membership probability (<0.60) of any phenotype. Class membership certainty was highest in persistent and never/infrequent, and lowest in late-onset wheeze (with 51% of participants having membership probabilities <0.80). Individual wheezing patterns were particularly heterogeneous in late-onset wheeze, whereas many children assigned to early onset preschool remitting class reported wheezing at later time points. Conclusions: All wheeze phenotypes had significantly diminished lung function in school-age children, suggesting that the notion that early life episodic wheeze has a benign prognosis may not be true for a proportion of transient wheezers. We observed considerable within-phenotype heterogeneity in individual wheezing patterns.

Practical implementation of artificial intelligence algorithms in pulmonary auscultation examination.

Lung auscultation is an important part of a physical examination. However, its biggest drawback is its subjectivity. The results depend on the experience and ability of the doctor to perceive and distinguish pathologies in sounds heard via a stethoscope. This paper investigates a new method of automatic sound analysis based on neural networks (NNs), which has been implemented in a system that uses an electronic stethoscope for capturing respiratory sounds. It allows the detection of auscultatory sounds in four classes: wheezes, rhonchi, and fine and coarse crackles. In the blind test, a group of 522 auscultatory sounds from 50 pediatric patients were presented, and the results provided by a group of doctors and an artificial intelligence (AI) algorithm developed by the authors were compared. The gathered data show that machine learning (ML)-based analysis is more efficient in detecting all four types of phenomena, which is reflected in high values of recall (also called as sensitivity) and F1-score.Conclusions: The obtained results suggest that the implementation of automatic sound analysis based on NNs can significantly improve the efficiency of this form of examination, leading to a minimization of the number of errors made in the interpretation of auscultation sounds. What is Known: • Auscultation performance of average physician is very low. AI solutions presented in scientific literature are based on small data bases with isolated pathological sounds (which are far from real recordings) and mainly on leave-one-out validation method thus they are not reliable. What is New: • AI learning process was based on thousands of signals from real patients and a reliable description of recordings was based on multiple validation by physicians and acoustician resulting in practical and statistical prove of AI high performance.

Wheeze type classification using non-dyadic wavelet transform based optimal energy ratio technique.

BACKGROUND AND OBJECTIVE: Wheezes in pulmonary sounds are anomalies which are often associated with obstructive type of lung diseases. The previous works on wheeze-type classification focused mainly on using fixed time-frequency/scale resolution based on Fourier and wavelet transforms. The main contribution of the proposed method, in which the time-scale resolution can be tuned according to the signal of interest, is to discriminate monophonic and polyphonic wheezes with higher accuracy than previously suggested time and time-frequency/scale based methods. METHODS: An optimal Rational Dilation Wavelet Transform (RADWT) based peak energy ratio (PER) parameter selection method is proposed to discriminate wheeze types. Previously suggested Quartile Frequency Ratios, Mean Crossing Irregularity, Multiple Signal Classification, Mel-frequency Cepstrum and Dyadic Discrete Wavelet Transform approaches are also applied and the superiority of the proposed method is demonstrated in leave-one-out (LOO) and leave-one-subject-out (LOSO) cross validation schemes with support vector machine (SVM), k nearest neighbor (k-NN) and extreme learning machine (ELM) classifiers. RESULTS: The results show that the proposed RADWT based method outperforms the state-of-the-art time, frequency, time-frequency and time-scale domain approaches for all classifiers in both LOO and LOSO cross validation settings. The highest accuracy values are obtained as 86% and 82.9% in LOO and LOSO respectively when the proposed PER features are fed into SVM. CONCLUSIONS: It is concluded that time and frequency domain characteristics of wheezes are not steady and hence, tunable time-scale representations are more successful in discriminating polyphonic and monophonic wheezes when compared with conventional fixed resolution representations.

Respiratory sound analysis in the era of evidence-based medicine and the world of medicine 2.0.

OBJECTIVE: This paper describes the state of the art, scientific publications, and ongoing research related to the methods of analysis of respiratory sounds. METHODS AND MATERIAL: Narrative review of the current medical and technological literature using Pubmed and personal experience. RESULTS: We outline the various techniques that are currently being used to collect auscultation sounds and provide a physical description of known pathological sounds for which automatic detection tools have been developed. Modern tools are based on artificial intelligence and techniques such as artificial neural networks, fuzzy systems, and genetic algorithms. CONCLUSION: The next step will consist of finding new markers to increase the efficiency of decision-aiding algorithms and tools.

Asthmatic Wheeze Detection From Compressively Sensed Respiratory Sound Spectra.

Quantification of wheezing by a sensor system consisting of a wearable wireless acoustic sensor and smartphone performing respiratory sound classification may contribute to the diagnosis, long-term control, and lowering treatment costs of asthma. In such battery-powered sensor system, compressive sensing (CS) was verified as a method for simultaneously cutting down power cost of signal acquisition, compression, and communication on the wearable sensor. Matching real-time CS reconstruction algorithms, such as orthogonal matching pursuit (OMP), have been demonstrated on the smartphone. However, their lossy performance limits the accuracy of wheeze detection from CS-recovered short-term Fourier spectra (STFT), when using existing respiratory sound classification algorithms. Thus, here we present a novel, robust algorithm tailored specifically for wheeze detection from the CS-recovered STFT. The proposed algorithm identifies occurrence and tracks multiple individual wheeze frequency lines using hidden Markov model. The algorithm yields 89.34% of sensitivity, 96.28% of specificity, and 94.91% of accuracy on Nyquist-rate sampled respiratory sounds STFT. It enables for less than 2% loss of classification accuracy when operating over STFT reconstructed by OMP, at the signal compression ratio of up to 4 $\times$ (classification from only 25% signal samples). It features execution speed comparable to referent algorithms, and offers good prospects for parallelism.

Lung sounds classification using convolutional neural networks.

Lung sounds convey relevant information related to pulmonary disorders, and to evaluate patients with pulmonary conditions, the physician or the doctor uses the traditional auscultation technique. However, this technique suffers from limitations. For example, if the physician is not well trained, this may lead to a wrong diagnosis. Moreover, lung sounds are non-stationary, complicating the tasks of analysis, recognition, and distinction. This is why developing automatic recognition systems can help to deal with these limitations. In this paper, we compare three machine learning approaches for lung sounds classification. The first two approaches are based on the extraction of a set of handcrafted features trained by three different classifiers (support vector machines, k-nearest neighbor, and Gaussian mixture models) while the third approach is based on the design of convolutional neural networks (CNN). In the first approach, we extracted the 12 MFCC coefficients from the audio files then calculated six MFCCs statistics. We also experimented normalization using zero mean and unity variance to enhance accuracy. In the second approach, the local binary pattern (LBP) features are extracted from the visual representation of the audio files (spectrograms). The features are normalized using whitening. The dataset used in this work consists of seven classes (normal, coarse crackle, fine crackle, monophonic wheeze, polyphonic wheeze, squawk, and stridor). We have also experimentally tested dataset augmentation techniques on the spectrograms to enhance the ultimate accuracy of the CNN. The results show that CNN outperformed the handcrafted feature based classifiers.

Multichannel lung sound analysis for asthma detection.

BACKGROUND AND OBJECTIVE: Lung sound signals convey valuable information of the lung status. Auscultation is an effective technique to appreciate the condition of the respiratory system using lung sound signals. The prior works on asthma detection from lung sound signals rely on the presence of wheeze. In this paper, we have classified normal and asthmatic subjects using advanced signal processing of posterior lung sound signals, even in the absence of wheeze. METHODS: We collected lung sounds of 60 subjects (30 normal and 30 asthma) using a novel 4-channel data acquisition system from four different positions over the posterior chest, as suggested by the pulmonologist. A spectral subband based feature extraction scheme is proposed that works with artificial neural network (ANN) and support vector machine (SVM) classifiers for the multichannel signal. The power spectral density (PSD) is estimated from extracted lung sound cycle using Welch's method, which then decomposed into uniform subbands. A set of statistical features is computed from each subband and applied to ANN and SVM classifiers to classify normal and asthmatic subjects. RESULTS: In the first part of this study, the performances of each individual channel and four channels together are evaluated where the combined channel performance is found superior to that of individual channels. Next, the performances of all possible combinations of the channels are investigated and the best classification accuracies of 89.2( ±â€¯3.87)% and 93.3( ±â€¯3.10)% are achieved for 2-channel and 3-channel combinations in ANN and SVM classifiers, respectively. CONCLUSIONS: The proposed multichannel asthma detection method where the presence of wheeze in lung sound is not a necessary requirement, outperforms commonly used lung sound classification methods in this field and provides significant relative improvement. The channel combination study gives insight into the contribution of respective lung sound collection areas and their combinations in asthma detection.