Heart Sound Analysis Using MFCC and Time Frequency Distribution

Kamarulafizam, I.; Salleh, Sheikh Hussain Shaikh; Najeb, J. M.; Ariff, A. K.; Chowdhury, Amor

doi:10.1007/978-3-540-36841-0_225

Cited by 17 publications

(10 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…9 frequency bands × 4 states = 36 features). Additionally, 13 mel-frequency cepstral coefficient (MFCC) [6] were extracted from each state and each cardiac cycle. The mean of MFCCs across different cardiac cycles from the same heart sound recording was used as MFCC features (i.e.…”

Section: Frequency-domain Featuresmentioning

confidence: 99%

Ensemble of Feature:based and Deep learning:based Classifiers for Detection of Abnormal Heart Sounds

Potes¹,

Parvaneh²,

Rahman³

et al. 2016

2016 Computing in Cardiology Conference (CinC)

192

161

View full text Add to dashboard Cite

The goal of the 2016 PhysioNet/CinC Challenge is the development of an algorithm to classify normal/abnormal heart sounds. A total of 124 time-frequency features were extracted from the phonocardiogram (PCG) and input to a variant of the AdaBoost classifier. A second classifier using convolutional neural network (CNN) was trained using PCGs cardiac cycles decomposed into four frequency bands. The final decision rule to classify normal/abnormal heart sounds was based on an ensemble of classifiers combining the outputs of AdaBoost and the CNN. The algorithm was trained on a training dataset (normal= 2575, abnormal= 665) and evaluated on a blind test dataset. Our classifier ensemble approach obtained the highest score of the competition with a sensitivity, specificity, and overall score of 0.9424, 0.7781, and 0.8602, respectively. IntroductionHeart auscultation is the primary tool for screening and diagnosis in primary health care [1]. Availability of digital stethoscopes and mobile devices provides clinicians an opportunity to record and analyze heart sounds (PCG) for diagnostic purposes. The goal of the 2016 PhysioNet/CinC Challenge is the development of algorithms to classify normal/abnormal heart sound recordings [2]. We proposed an ensemble of a feature-based classifier and a deep learningbased classifier to boost the classification performance of heart sounds. Method and MaterialA block diagram of the proposed approach to classify normal/abnormal PCG is shown in Fig. 1. Challenge DatabaseThe challenge database provided PCG recordings of healthy subjects and pathological patients collected at either a clinical or non-clinical environment. Details about the challenge dataset can be found in [2]. For algorithm development, in-house training and test sets were generated by randomly taking 80% and 20% of the records from each database, while keeping the same prevalence of abnormal classes. In-house training set was used for training and cross-validation of different models, and in-house test set was used for evaluation of the classification performance independently from the blind test dataset. Pre-processingEach PCG was resampled to 1000 Hz, band-pass filtered between 25 Hz and 400 Hz, and then pre-processed to remove any spikes in the PCG [3]. Furthermore, preprocessed PCGs were segmented into four heart sound states using a segmentation method proposed by Springer et al. [4]. Each PCG is comprised of more than one cardiac cycle (beat), and each beat is comprised of four heart sound states (i.e. S1, systole, S2, and diastole). Feature-based ApproachIn this approach, a variant of AdaBoost classifier [5] was trained for classification of normal/abnormal PCGs using time and frequency-domain features. Time-domain FeaturesMean and standard deviation (SD) of the following parameters were used as time-domain features (36 features): 1. PCG intervals: RR intervals, S1 intervals, S2 intervals, systolic intervals, diastolic intervals, ratio of systolic interval to RR interval of each heart beat, ratio of diastolic...

show abstract

Section: Frequency-domain Featuresmentioning

confidence: 99%

Ensemble of Feature:based and Deep learning:based Classifiers for Detection of Abnormal Heart Sounds

Potes¹,

Parvaneh²,

Rahman³

et al. 2016

2016 Computing in Cardiology Conference (CinC)

192

161

View full text Add to dashboard Cite

show abstract

“…Hal ini dikarenakan metode Mel Frequency Cepstral Coefficient (MFCC) mengadaptasi dari prinsipprinsip pendengaran manusia. Metode ini juga digunakan untuk menganalisa bagaimana Fourier Transform untuk mengekstrak komponen frekuensi dari sinyal dalam domai waktu (Kamarulafizam et al, 2007).…”

Section: Mel Frequency Cepstral Coefficient (Mfcc)unclassified

Klasifikasi Gender Berdasarkan Suara Dengan Naive Bayes Dan Mel Frequency Cepstral Coefficient

Safriadi¹,

Rahmadani

2020

VCT

View full text Add to dashboard Cite

Bagi manusia mengenali suara merupakan hal yang mudah, dengan cara mendengarkan dengan seksama dan manusia mempunyai kecerdasan dalam mengenali pola suara. Berbeda dengan komputer, proses pengenalan suara merupakan proses yang sulit, hal ini dikarenakan komputer memerlukan suatu mekanisme yang standar dan logis dalam mengenali pola suara. Dengan metode Mel Frequency Cepstral Coefficient (MFCC) memiliki peran penting dalam menentukan karakteristik dari sebuah suara. Metode ini sering digunakan untuk verifikasi suara, pengenalan suara, deteksi emosi dari suara. Untuk melakukan klasifikasi pada penelitian ini menggunakan metode Naïve Bayes. Metode Naive Bayes merupakan salah satu metode klasifikasi, yang mana proses klasifikasi pada metode naïve bayes berdasarkan dari probabilitas dari data sebagai bukti dalam probalitas. Model yang digunakan pada metode Naive Bayes adalah model atribut independent. Dalam penelitian ini, data suara yang digunakan pada penelitian ini berupa data suara yang direkam mengunkan perekam suara dengan durasi rekaman suara maksimal 30 detik. Tingkat keberhasilan dalam penelitian ini sebesar 87%. Hal ini berdasarkan dari jumlah data pengujian 100 sampel, yang benar diklasifikasi sebanyak 87 sampel data sedangkan yang salah diklasifikasi sebanyak 13 data sampel suara. For humans, recognizing sounds is an easy thing, by listening carefully and understandingly to what is spoken and humans have intelligence in recognizing sound patterns. Unlike computers, the speech recognition process is a difficult process, this is because the computer requires a standard and logical mechanism to recognize sound patterns. With Mel Frequency Cepstral Coefficient (MFCC) method has an important role in determining the characteristics of a sound. This method is often used for verification of voice, speech recognition, emotion detection of voice. To perform the classification in this study using Naïve Bayes method. The Naive Bayes method is a classification method. In which the classification process in the naïve Bayes method is based on the probability of the data as evidence in probability. The model used in the Naive Bayes method is the independent attribute model. The accuracy rate in this research was 87%. It is based on the amount of data testing 100 samples, the true classified as 87 samples of data while false classified as 13 sample data.

show abstract

“…However, the system was trained on a single sample for each disease using different heart beat cycles. Heart sound analysis using time-frequency representations has also been common including recent uses of MFCC to heart sounds [5].…”

Section: Related Workmentioning

confidence: 99%

“…The MFCC method was chosen as it has been the most successful of the audio analysis approaches and has recently been used to classify heart diseases [5]. We implemented a version of MFCC in which we divided the raw audio signals (no periodicity detection) into short-time segments of 400 samples and MFCC coefficients were then extracted from each segment using the short-time Fourier transform (STFT).…”

Section: Comparison With Mfccmentioning

confidence: 99%

Shape-Based Retrieval of Heart Sounds for Disease Similarity Detection

Syeda-Mahmood

Wang

2008

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. Retrieval of similar heart sounds from a sound database has applications in physician training, diagnostic screening, and decision support. In this paper, we exploit a visual rendering of heart sounds and model the morphological variations of audio envelopes through a constrained non-rigid translation transform. Similar heart sounds are then retrieved by recovering the corresponding alignment transform using a variant of shape-based dynamic time warping. Results of similar heart sound retrieval are demonstrated for various diseases on a large database of heart sounds.

show abstract

Heart Sound Analysis Using MFCC and Time Frequency Distribution

Cited by 17 publications

References 11 publications

Ensemble of Feature:based and Deep learning:based Classifiers for Detection of Abnormal Heart Sounds

Ensemble of Feature:based and Deep learning:based Classifiers for Detection of Abnormal Heart Sounds

Klasifikasi Gender Berdasarkan Suara Dengan Naive Bayes Dan Mel Frequency Cepstral Coefficient

Shape-Based Retrieval of Heart Sounds for Disease Similarity Detection

Contact Info

Product

Resources

About