Respiratory sound classification using cepstral features and support vector machine

In the field of medicine, with the introduction of computer systems that can collect and analyze massive amounts of data, many non-invasive diagnostic methods are being developed for a variety of conditions. In this study, our aim is to develop a non-invasive method of classifying respiratory sounds that are recorded by an electronic stethoscope and the audio recording software that uses various machine learning algorithms.In order to store respiratory sounds on a computer, we developed a cost-effective and easy-to-use electronic stethoscope that can be used with any device. Using this device, we recorded 17,930 lung sounds from 1630 subjects.We employed two types of machine learning algorithms; mel frequency cepstral coefficient (MFCC) features in a support vector machine (SVM) and spectrogram images in the convolutional neural network (CNN). Since using MFCC features with a SVM algorithm is a generally accepted classification method for audio, we utilized its results to benchmark the CNN algorithm. We prepared four data sets for each CNN and SVM algorithm to classify respiratory audio: (1) healthy versus pathological classification; (2) rale, rhonchus, and normal sound classification; (3) singular respiratory sound type classification; and (4) audio type classification with all sound types. Accuracy results of the experiments were; (1) CNN 86%, SVM 86%, (2) CNN 76%, SVM 75%, (3) CNN 80%, SVM 80%, and (4) CNN 62%, SVM 62%, respectively.As a result, we found out that spectrogram image classification with CNN algorithm works as well as the SVM algorithm, and given the large amount of data, CNN and SVM machine learning algorithms can accurately classify and pre-diagnose respiratory audio.

This work addresses the problem of lung sound classification, in particular, the problem of distinguishing between wheeze and normal sounds. Wheezing sound detection is an important step to associate lung sounds with an abnormal state of the respiratory system, usually associated with tuberculosis or another chronic obstructive pulmonary diseases (COPD). The paper presents an approach for automatic lung sound classification, which uses different state-of-the-art sound features in combination with a C-weighted support vector machine (SVM) classifier that works better for unbalanced data. Feature extraction methods used here are commonly applied in speech recognition and related problems thanks to the fact that they capture the most informative spectral content from the original signals. The evaluated methods were: Fourier transform (FT), wavelet decomposition using Wavelet Packet Transform bank of filters (WPT) and Mel Frequency Cepstral Coefficients (MFCC). For comparison, we evaluated and contrasted the proposed approach against previous works using different combination of features and/or classifiers. The different methods were evaluated on a set of lung sounds including normal and wheezing sounds. A leave-two-out per-case cross-validation approach was used, which, in each fold, chooses as validation set a couple of cases, one including normal sounds and the other including wheezing sounds. Experimental results were reported in terms of traditional classification performance measures: sensitivity, specificity and balanced accuracy. Our best results using the suggested approach, C-weighted SVM and MFCC, achieve a 82.1% of balanced accuracy obtaining the best result for this problem until now. These results suggest that supervised classifiers based on kernel methods are able to learn better models for this challenging classification problem even using the same feature extraction methods.

Many deep learning–based computerized respiratory sound analysis methods have previously been developed. However, these studies focus on either lung sound only or tracheal sound only. The effectiveness of using a lung sound analysis algorithm on tracheal sound and vice versa is rarely reported. Furthermore, no one knows whether using lung and tracheal sounds together in training a deep learning-based respiratory sound analysis model is beneficial. In this study, we first constructed a tracheal sound database, HF_Tracheal_V1, containing 10,448 15-s tracheal sound recordings, 21,741 inhalation labels, 15,858 exhalation labels, and 6414 continuous adventitious sound (CAS) labels collected from 227 participants undergoing a diagnostic/surgical procedure under monitored anesthesia care. HF_Tracheal_V1 and our previously built lung sound database, HF_Lung_V2, were either combined (mixed set), used one after the other (domain adaptation), or used alone to train convolutional neural network bidirectional gate recurrent unit models for inhalation, exhalation, and CAS detection in lung and tracheal sounds. The results revealed that the models trained using lung sound alone performed poorly in tracheal sound analysis and vice versa. However, mixed set training or domain adaptation improved the performance for 1) inhalation and exhalation detection in lung sounds and 2) inhalation, exhalation, and CAS detection in tracheal sounds compared to positive controls (the models trained using lung sound alone and used in lung sound analysis and vice versa). In particular, the model trained on the mixed set had great flexibility to serve two purposes, lung and tracheal sound analyses, at the same time.

Classification of respiratory sounds between normal and abnormal is very crucial for screening and diagnosis purposes. Lung associated diseases can be detected through this technique. With the advancement of computerized auscultation technology, the adventitious sounds such as crackles can be detected and therefore diagnostic test can be performed earlier. In this paper, Mel-frequency Cepstral Coefficient (MFCC) is used to extract features from normal and crackles respiratory sounds. By using statistical computation such as mean and standard deviation (SD) of cepstral based coefficients it can differentiate between crackles and normal sounds. The result shows that the first three statistical values of SD of coefficients provide distinctive feature between normal and crackles respiratory sounds. Hence, MFCCs can be used as feature extraction method of respiratory sounds to classify between normal and crackles as screening and diagnostic tool.

We designed a computer-based respiratory sound analysis system to identify pediatric normal lung sound. To verify the validity of the computer-based respiratory sound analysis system. First we downloaded the standard lung sounds from the network database (website: http: //www.easyauscultation.com/lung-sounds-reference-guide) and recorded 3 samples of abnormal loud sound (rhonchi, wheeze and crackles) from three patients of The Department of Pediatrics, the First Affiliated Hospital of Xiamen University. We regarded such lung sounds as"reference lung sounds". The"test lung sounds"were recorded from 29 children form Kindergarten of Xiamen University. we recorded lung sound by portable electronic stethoscope and valid lung sounds were selected by manual identification. We introduced Mel-frequency cepstral coefficient (MFCC) to extract lung sound features and dynamic time warping (DTW) for signal classification. We had 39 standard lung sounds, recorded 58 test lung sounds. This computer-based respiratory sound analysis system was carried out in 58 lung sound recognition, correct identification of 52 times, error identification 6 times. Accuracy was 89.7%. Based on MFCC and DTW, our computer-based respiratory sound analysis system can effectively identify healthy lung sounds of children (accuracy can reach 89.7%), fully embodies the reliability of the lung sounds analysis system.

A boosted SVM classifier trained by incremental learning and decremental unlearning approach

This paper presents a new approach for the acoustic classification of frogs' calls using a novel fusion of features: Mel Frequency Cepstral Coefficients (MFCCs), Shannon entropy and syllable duration. First, the audio recordings of different frogs' species are segmented in syllables. For each syllable, each feature is extracted and the cepstral features (MFCC) are computed and evaluated separately as in previous works. Finally, the data fusion is used to train a multiclass Support Vector Machine (SVM) classifier. In our experiment, the results show that our novel feature fusion increase the classification accuracy; achieving an average of 94.21% ± 8,04 in 18 frog's species.

Background/Objectives: Identifying and classifying abnormal lung sounds is essential for diagnosing patients with respiratory disorders. In particular, the simultaneous recording of auscultation signals from multiple clinically relevant positions offers greater diagnostic potential compared to traditional single-channel measurements. This study aims to improve the accuracy of respiratory sound classification by leveraging multichannel signals and capturing positional characteristics from multiple sites in the same patient. Methods: We evaluated the performance of respiratory sound classification using multichannel lung sound data with a deep learning model that combines a convolutional neural network (CNN) and long short-term memory (LSTM), based on mel-frequency cepstral coefficients (MFCCs). We analyzed the impact of the number and placement of channels on classification performance. Results: The results demonstrated that using four-channel recordings improved accuracy, sensitivity, specificity, precision, and F1-score by approximately 1.11, 1.15, 1.05, 1.08, and 1.13 times, respectively, compared to using three, two, or single-channel recordings. Conclusions: This study confirms that multichannel data capture a richer set of features corresponding to various respiratory sound characteristics, leading to significantly improved classification performance. The proposed method holds promise for enhancing sound classification accuracy not only in clinical applications but also in broader domains such as speech and audio processing.

The aim of the study was to inspect the acoustic properties and sleep characteristics of a preapneic snoring sound. The feasibility of forecasting upcoming respiratory events by snoring sound was also investigated. Participants with habitual snoring or a heavy breathing sound during sleep were recruited consecutively. Polysomnography was conducted, and snoring-related breathing sound was recorded simultaneously. Acoustic features and sleep features were extracted from 30-second samples, and a machine learning algorithm was used to establish 2 prediction models. A total of 74 eligible participants were included. Model 1, tested by 5-fold cross-validation, achieved an accuracy of 0.92 and an area under the curve of 0.94 for respiratory event prediction. Model 2, with acoustic features and sleep information tested by Leave-One-Out cross-validation, had an accuracy of 0.78 and an area under the curve of 0.80. Sleep position was found to be the most important among all sleep features contributing to the performance of the 2 models. Preapneic sound presented unique acoustic characteristics, and snoring-related breathing sound could be deployed as a real-time apneic event predictor. The models, combined with sleep information, serve as a promising tool for an early warning system to forecast apneic events. Wang B, Yi X, Gao J, etal. Real-time prediction of upcoming respiratory events via machine learning using snoring sound signal. J Clin Sleep Med. 2021;17(9):1777-1784.

Respiratory sound classification has emerged as a promising non-invasive and scalable tool for the early detection of respiratory disorders. While most previous studies have relied on a single feature extraction method-such as Melfrequency cepstral coefficients (MFCC), Mel-spectrograms, or Short-Time Fourier Transform (STFT)-this study provides a comprehensive comparative analysis of these three approaches, evaluated individually, in pairwise combinations, and in a combined three-method configuration. Experiments are conducted on the SPRSound dataset, a pediatric respiratory sound database comprising 6,656 annotated events from seven unbalanced classes. Three architectures are assessed under identical preprocessing and augmentation strategies: a custom convolutional neural network (CNN), and two pre-trained models (VGG16 and InceptionV3) fine-tuned via transfer learning. Results show that STFT consistently delivers the highest performance for CNN and VGG16 models, while MFCC achieves the best accuracy with InceptionV3. Specifically, VGG16 with STFT attained 93.43% accuracy (score: 0.9545), whereas InceptionV3 with MFCC achieved the top performance within its architecture. These findings highlight the importance of aligning feature extraction techniques with model architecture and provide a systematic benchmark for SPRSound-based respiratory sound classification.

This paper presents a comparison between several features normalization methods, and a comparison between different types of Gaussian Mixture Model (GMM) based supervectors normalizations for robust Speaker Verification. We implemented the methods of normalizations as a part of speaker verification system using Support Vector Machine (SVM) classifier and GMM-based supervectors. When implementing the speaker recognition system, we used Mel Frequency Cepstral Coefficients (MFCC) feature extraction. A valid question is which features normalization to use, if any. We examine the most common methods of feature normalizations, such as: Feature Warping mapping, and Cepstral Mean Subtraction (CMS) normalization with and without variance normalization. These methods were compared to features without normalization at all, and to a basic [-1, 1] normalization. In addition, we applied few types of normalizations to the GMM-mean supervectors, in order to improve the performance of the SVM classifier. All comparisons of the speaker verification system had been done in terms of DET curve, EER (Equal Error Rate) and Min. DCF. The best results we achieved were on combined supervector normalizations of Universal Background Model (UBM) Standard Deviation (STD) and [-1, 1] normalization. The type of the MFCC normalization has no big influence on the verification performance. The best results were: EER about 5.0% and MIN. DCF of 0.02.

In order to improve the performance of Support Vector Machine (SVM) classifier for imbalanced data, this paper proposes dynamic cost-sensitive SVM classifier based on chaos particle swarm optimization (CPDC_SVM). Firstly, this paper introduces dynamic cost-sensitive thought to SVM classifier, and gives the method for structuring dynamic cost and cost-sensitive SVM model. Secondly, we propose the evaluation methodology performance for classifier, and adopts decimal base to code the particles. At last, chaos thought is introduced in particle swarm optimization algorithm, and the Algorithm of the dynamic cost-sensitive SVM classifier is given, which improves convergent speed and accuracy of particle swarm optimization, and can optimize dynamic cost-sensitive SVM well, so CPDC_SVM adds effectively the convergence speed and accuracy for the particle swarm optimization algorithm. Experimental results show CPDC_SVM has higher precision than traditional SVM classifier, and dynamic cost and chaos particle swarm optimization can improve the performance for classifier.

Lung sound classification using cepstral-based statistical features

This paper presents and explores a robust deep learning framework for auscultation analysis. This aims to classify anomalies in respiratory cycles and detect diseases, from respiratory sound recordings. The framework begins with front-end feature extraction that transforms input sound into a spectrogram representation. Then, a back-end deep learning network is used to classify the spectrogram features into categories of respiratory anomaly cycles or diseases. Experiments, conducted over the ICBHI benchmark dataset of respiratory sounds, confirm three main contributions towards respiratory-sound analysis. Firstly, we carry out an extensive exploration of the effect of spectrogram types, spectral-time resolution, overlapping/non-overlapping windows, and data augmentation on final prediction accuracy. This leads us to propose a novel deep learning system, built on the proposed framework, which outperforms current state-of-the-art methods. Finally, we apply a Teacher-Student scheme to achieve a trade-off between model performance and model complexity which holds promise for building real-time applications.

A query image based scene/image retrieval system is a system that analyzes the properties of a query image and identifies the class in which the image belongs and retrieves a number of images which are most alike and relevant to the query image. A scene/image classifier provides the first stage for this system. Scene classification is the process that analyzes the properties of various image features of a scene and organizes the scenes into categories. Programmed techniques for correlating scenes with labels have a high potential for improving the performance of computer vision applications such as searching, retrieval and object recognition. The scene category system uses the feature vectors extracted from image data to classify test images. We have implemented a k-nearest neighbour (kNN) classifier which uses gray scale and Scale Invariant Feature Transform (SIFT) features as input and used support vector machine (SVM) classifier that uses GIST features as input. kNN classifier is used, as it has an advantage of working with a large amount of data and SVM is robust even when the training sample has some bias. The performance measure of the classifier was evaluated using confusion matrix and classification accuracy. We considered MIT-8 scene category, leaves, Caltech-20 and deer data sets for our experiments. The classification accuracy was found to be good. Results show that classification accuracy is increased significantly for classification using GIST features in SVM classifiers. We are exploring methods to obtain optimal performance of classifiers.