Classification of heart diseases from ECG signals using wavelet transform and kNN classifier

Saini, Ridhi; Bindal, Namita; Bansal, Puneet

doi:10.1109/ccaa.2015.7148561

Cited by 39 publications

(17 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The usage of wavelet-compressed data from the St. Petersburg Institute of Cardiological Technics 12lead arrhythmia database was our first step. In general, the MIT-BIH arrhythmia database is used by most research that is performed on ECG signals [1][2][3][4][5][6][7][8][9][10][11][12]14,16,19,[21][22][23]25,26], and classification is executed on a single channel [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][21][22][23][24][25][26][27]. Our research was performed on 12 channels.…”

Section: Discussionmentioning

confidence: 99%

“…The SVM [4,5,8,14,16,18,21,22,24] is another popular technique for classification of arrhythmia. Random forest [9], linear classifier [10], morphology consistency evaluation [13], cluster and centroid identification [15], linear discriminant analysis [17], threshold based classifier [20], KNN [7,23], and naive Bayes [26] are the other classification methods that are used in related works. The results obtained from 12-lead ECG recordings using bagged trees classification showed that our performance was as successful as others and more stable.…”

Section: Discussionmentioning

confidence: 99%

“…Saini et al [23] classified ten heart diseases using the K-nearest neighbors (KNN) classifier. The classification was done using wavelet-transformed ECG signals and the original ECG signals.…”

Section: Llamedo and Martínezmentioning

confidence: 99%

See 2 more Smart Citations

Bagged tree classi cation of arrhythmia using wavelets for denoising, compression, and feature extraction

Tomak¹,

Kayıkçıoğlu²

2018

Turk J Elec Eng & Comp Sci

View full text Add to dashboard Cite

Arrhythmia, also known as dysrhythmia, is a condition involving an irregular heartbeat. A problem in the heart may cause problems in other organs, and as time passes, this will lead to more severe problems. Arrhythmia must be detected at an early stage to prevent such a problem occurring in the heart. Detection of arrhythmia from an electrocardiogram is an easy method that does not need much equipment and does not harm the patient. The purpose of this research is to find a faster and more accurate system to classify nine classes of arrhythmia. The St. Petersburg Institute of Cardiological Technics 12-lead arrhythmia database was used for training and testing. Data were compressed and preprocessed (denoising, trend elimination, baseline correction, and normalization) before being sent to the system for feature calculation. The wavelet coefficients that displayed the most significant effect on classification were chosen and used as features. Standard deviation and variance were also added to the feature set. Later, principal component analysis (PCA) was used to reduce the number of features further. After deciding the features, the performance of the basic classification methods and spiking neural network was checked to determine whether there was a better classifier to be used for our research. Tenfold cross-validation was applied to the training dataset. Bagged trees were found to produce better results. The classifiers' performance was tested by sensitivity, specificity, and accuracy.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Bagged tree classi cation of arrhythmia using wavelets for denoising, compression, and feature extraction

Tomak¹,

Kayıkçıoğlu²

2018

Turk J Elec Eng & Comp Sci

View full text Add to dashboard Cite

show abstract

“…Although feature methodology tries to quantify ECG in a discriminative way, the heuristics of hand-crafting process and the complicated variation of signals unavoidably impede the performance enhancement. Therefore, pattern analysis and/or machine learning techniques play an important role in further improving the classification performance based on the extracted features [ 16 , 17 , 18 , 19 , 20 , 21 , 22 ].…”

Section: Introductionmentioning

confidence: 99%

Set-Based Discriminative Measure for Electrocardiogram Beat Classification

Qin

2017

Sensors

View full text Add to dashboard Cite

Abstract:Computer aided diagnosis systems can help to reduce the high mortality rate among cardiac patients. Automatical classification of electrocardiogram (ECG) beats plays an important role in such systems, but this issue is challenging because of the complexities of ECG signals. In literature, feature designing has been broadly-studied. However, such methodology is inevitably limited by the heuristics of hand-crafting process and the challenge of signals themselves. To address it, we treat the problem of ECG beat classification from the metric and measurement perspective. We propose a novel approach, named "Set-Based Discriminative Measure", which first learns a discriminative metric space to ensure that intra-class distances are smaller than inter-class distances for ECG features in a global way, and then measures a new set-based dissimilarity in such learned space to cope with the local variation of samples. Experimental results have demonstrated the advantage of this approach in terms of effectiveness, robustness, and flexibility based on ECG beats from the MIT-BIH Arrhythmia Database.

show abstract

“…Por ejemplo, para la detección y clasificación de etapas de sueño a partir del ECG (Yilmaz et al, 2010) y para la detección de períodos de apnea en combinación con AdaBoost Bootstrap (Kao et al, 2012). También se ha usado KNN para diagnosticar enfermedades del corazón de forma automática (Saini et al, 2015).…”

Section: Introductionunclassified

Detección de Fibrilación Ventricular Mediante Tiempo-Frecuencia y Clasificador KNN sin Extracción de Parámetros

Mjahad

Muñoz

Mompean

et al. 2017

Rev. iberoam. autom. inform. ind.

View full text Add to dashboard Cite

ResumenEste trabajo propone la detección de FV y su discriminación de TV y otros ritmos cardiacos basándose en la representación tiempo-frecuencia del ECG y su conversión en imágen como entrada a un clasificador de vecinos más cercanos (KNN) sin necesidad de extracción de parámetros adicionales. Tres variantes de datos de entrada al clasificador son evaluados. Los resultados clasifican la señal en cuatro clases diferentes: 'Normal' para latidos con ritmo sinusal, 'FV' para fibrilación ventricular, 'TV' para taquicardia ventricular y 'Otros' para el resto de ritmos. Los resultados para detección de FV mostraron 88,27 % de sensibilidad y 98,22 % de especificidad para la entrada de imagen equivalente reducida que es la más rápida computacionalmente a pesar de obtener resultados de clasificación ligeramente inferiores a las representaciones no reducidas. En el caso de TV, se alcanzó un 88,31 % de sensibilidad y 98,80 % de especificidad, un 98,14 % de sensibilidad y 96,82 % de especificidad para ritmo sinusal normal y 96,91 % de sensibilidad con 99,06 % de especificidad para la clase 'Otros'. Finalmente, se realiza una comparación con otros algoritmos.Palabras Clave: Sistemas biomédicos, Señales Electrocardiográficas, Representación tiempo-frecuencia, Señales no estacionarias, Análisis de imágenes, Clasificación Ventricular Fibrillation detection using time-frequency and the KNN classifier without parameter extraction. AbstractThis work describes new techniques to improve VF detection and its separation from Ventricular Tachycarida (VT) and other rhythms. It is based on time-frequency representation of the ECG and its use as input in an automatic classifier (K-nearest neighbours -KNN) without any further signal parameter extraction or additional characteristics. For comparison purposes, three time-frequency variants are analysed: pseudo Wigner-Ville representation (RTF), grey-scale image obtained from RTF (IRTF), and reduced image from IRTF (reduced IRTF). Four types of rhythms (classes) are defined: 'Normal' for sinus rhythm, 'VT' for ventricular tachycardia, 'VF' for ventricular fibrillation and 'Others' for the rest of rhythms. Classification results for VF detection in case of reduced IRTF are 88.27 % sensitivity and 98.22 % specificity. In case of VT, 88.31 % sensitivity and 98.80 % specificity is obtained, 98.14 % sensitivity and 96.82 % specificity for normal rhythms, and 96.91 % sensitivity and 99.06 % specificity for other rhythms. Finally, results are compared with other authors.

show abstract

Classification of heart diseases from ECG signals using wavelet transform and kNN classifier

Cited by 39 publications

References 9 publications

Bagged tree classi cation of arrhythmia using wavelets for denoising, compression, and feature extraction

Bagged tree classi cation of arrhythmia using wavelets for denoising, compression, and feature extraction

Set-Based Discriminative Measure for Electrocardiogram Beat Classification

Detección de Fibrilación Ventricular Mediante Tiempo-Frecuencia y Clasificador KNN sin Extracción de Parámetros

Contact Info

Product

Resources

About