Cardiac Arrhythmia Classification Using Hjorth Descriptors

  • Thaweesak YingthawornsukEmail author
  • Pawita Temsang
Conference paper
Part of the Advances in Intelligent Systems and Computing book series (AISC, volume 807)


The aim of this proposed study is to investigate the discriminant power of Hjorth Descriptor in classification of three categorized groups of subjects’ ECG measurement, which are Normal Sinus Rhythm (NSR), Atrial Fibrillation (AF) and Congestive Heart Failure (CHF). This feature has been previously employed to measure the healthiness in persons via their ECG recordings. The algorithm was designed and implemented to extract the Hjorth features and evaluate the performance of classification made on those features by comparing all classifications made among those three databases. Each categorized group included thirty subjects evenly and only three complete QRS complexes of each record in our databases were selected, segmented and extracted for their Hjorth descriptor estimators. In this work three different classifiers were selected, which are Least-Squares (LS), Maximum likelihood (ML) and Support Vector Machine (SVM) for performance evaluation and accuracy comparison. The experimental results from our study showed that the most effective classifier was found to be ML with a mean accuracy of 84.89%, SE of 88.82% and SP of 99.75%, as compared to LS which was found to be the second effective classifier with 88.22% accuracy, and finally SVM with 76.94%. These findings suggested that the promisingly dominant ECG based Hjorth descriptor is capable of class separation among cardiac arrhythmia patient groups.


Electrocardiogram Hjorth descriptor Classification 


  1. 1.
    Brosche, T.A.M.: The EKG Handbook. Jones & Bartlett Publisher (2010)Google Scholar
  2. 2.
    Bollmann, A., Roig, M., Castells, F., Laguna, P., Leif, S.: Principal component analysis in ECG signal processing. EURASIP J. Adv. Signal Process. 2007, 074580 (2007)Google Scholar
  3. 3.
    Xiao, Q.U., Wei, C.: ECG signal classification based on BPNN. In: 2011 International Conference on Electric Information and Control Engineering (2011)Google Scholar
  4. 4.
    Daamouche, A., Hamami, L., Alajlan, N., Melgani, F.: A wavelet optimization approach for ECG signal classification. Biomed. Signal Process. Control 7(4), 342–349 (2012)CrossRefGoogle Scholar
  5. 5.
    Hadiyoso, S., Rizal, A.: ECG Signal Classification using Higher-Order Complexity of Hjorth Descriptor. American Scientific Publishers (2015)Google Scholar
  6. 6.
    Rizal, A., Hadiyoso, S.: ECG Signal Classification Using Hjorth DescriptorGoogle Scholar
  7. 7., “ECG Database”.
  8. 8.
    Tompkins, W.J.: Electrocardiography. In: Tompkins, W.J. (ed.) Biomedical Digital Signal Processing. Prentice Hall, New Jersey, pp. 24–54 (2000)Google Scholar
  9. 9.
    Guyton, C: Textbook of Medical Physiology, 8th edn. Harcourt College Pub., October 1990Google Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  1. 1.Media Technology ProgramKing Mongkut’s University of Technology ThonburiBangkokThailand

Personalised recommendations