Control Systems and Computers

DOI: https://doi.org/10.15407/usim.2016.02.076

Upr. sist. maš., 2016, Issue 2 (262), pp. 76-84.

UDC 616.12-008.318.1

Fainzilberg Leonid S., Doctor of Technical Sciences, head of the department. International Research and Training Center for Information Technologies and Systems of the NAS and MES of Ukraine, Glushkov ave., 40, Kyiv, 03187, Ukraine, E-mail: fainzilberg@voliacable.com

Matushevych Nataliia A., student, National Technical University of Ukraine “Igor Sikorsky Kyiv Politechnic Institute”, Peremohy Ave 37, Kyiv, Ukraine, natalie.matushevych@gmail.com 

An Effective Method for Analysis of the Diagnostic Features Based on Noisy Electrocardiogram

Introduction. Computer algorithms of ECG processing often lead to errors on the stage of measuring the diagnostic features focused on the informative fragments of ECG. Therefore, the actual task is a construction of the methods that enhance the reliability of the signs assessment of the real ECG that was distorted by the noise.

The objective of the article is development of the ECG cycles approximation method by set of Gaussian functions, advancing on its basis a constructive algorithm ensuring an effective transition from the signal, observed on the conditions of the imposed external perturbations, to the diagnostic features system.

Methods. The proposed method is based on model descriptions of ECG cycle as the sum of asymmetrical Gaussian functions. The determination of the optimal parameter values of the function based on the criteria of the Least Squares Method.
The main electrocardiographic features are evaluated using the optimal parameters of the approximating function.

Results. As the Gaussian function is nonlinear by parameters, for the practical realization the analytical method of evaluation optimal parameters, it is proposed to modify the optimality criterion moving from approximation of the actual data to
their logarithms. For correct use of this approach, the original interval of approximation is narrowed to the desired value.

The model experiments have shown that the modification criterion and introduction of the restrictions on interval of approximation allows an analytical solutions for estimating the parameters of the approximating functions from noisy data with
the required accuracy. Even with a high noise level that in the experiments has reached 50% of the signal change range, mean square error of approximation was close to 1%.

Conclusions. The proposed method provides the evaluation of the major electrocardiographic features of real signal with the necessary accuracy for their proper interpretation.

Download full text! (Russian).

Keywords: electrocardiographic sign ECG, Gaussian functions, approximation, optimality criterion.

1. Jokic, S., Delic, V., Peric Z. et al., 2011. “Efficient ECG Modeling using Polynomial Functions”. Electronics and Electrical Engineering, 4(110), pp. 121–124.
   https://doi.org/10.5755/j01.eee.110.4.304
2. Riad, B., Amine, N., Jacques, L., 2004. “ECG Compression Using an Ensemble Polynomial Modeling: Comparison With the DCT Based Technique”. Cardiovascular Engineering, 4 (3), pp. 237–244.
   https://doi.org/10.1023/B:CARE.0000038780.96845.27
3. Kovacs, P., 2012. “ECG Signal Generator Based On Geometrical Features”. Annales Univ. Sci. Budapest., Sect. Comp., 37, pp. 247–260.
4. Vishnevskiy, V.V., Kalmykov, V.G., Romanenko, T.N., 2015. “Approksimatsiya odno-, dvu- i trekhmernykh dug krivykh parametricheskimi splaynami”. Matematychni mashyny i systemy, 4, pp. 57–64. (In Russian).
5. Vishnevskiy V.V., Rystsov I.K., Volzheva M.V. Iteratsionnyy algoritm postroyeniya krivoy Bez’ye po zadannym tochkam. Matematychni mashyny i systemy, 2004. 4. pp. 108–116. (In Russian).
6. McSharry Clifford, G., Tarasenko, L., Smith, L. A., 2003. “Dynamical Model For Generating Synthetic Electrocardiogram Signals”. IEEE Transaction on Biomedical Engin, 50 (3), pp. 289–294.
   https://doi.org/10.1109/TBME.2003.808805
7. Fainzilberg, L.S., Bekler, T.Yu., Glushauskene, G.A., 2011. “Mathematical Model for Generation of Artificial Electrocardiogram with Given Amplitude-Time Characteristics of Informative Fragments”. J. of Automation and Inform. Sci., 43 (9), pp. 20–33.
   https://doi.org/10.1615/JAutomatInfScien.v43.i9.20
8. Abramov, M.V., 2010. “Approksimatsii eksponentami vremennogo kardiologicheskogo ryada na osnove EKG”. Vestn. kibernetiki, 9, pp. 85–91. (In Russian).
9. Nikiforov, P.L., 1998. “Model’ elektrokardiograficheskogo signala na osnove sovokupnosti kolokol’nykh impul’sov”. Vestn. molodykh uchenykh. Seriya tekhn. nauk., 1, pp. 64–68. (In Russian).
10. Parvaneh, S., Pashna, M., 2007. “Electrocardiogram Synthesis Using a Gaussian Combination Model (GCM)”. Computers in Cardiology, 34, pp. 621−624.
    https://doi.org/10.1109/CIC.2007.4745562
11. Fainzilberg, L.S., 2015. “Generalized Method of Processing Cyclic Signals of Complex Form in Multidimension Space of Parameters”. J. of Automation and Inform. Sci., 47 (3), pp. 24–39.
    https://doi.org/10.1615/JAutomatInfScien.v47.i3.30
12. Murashko, V.V., Strutinskiy, A.V., 1991. Elektrokardiografiya. M.: Meditsina, 288 p. (In Russian).
13. Pan, J., Tompkins, W., 1985. “A Real-Time QRS Detection Algorithm”. IEEE Transaction on Biomed. Engin., 32 (3), pp. 230–236.
14. Strumillo, P., 2002. “Nested Median Filtering for Detecting T-Wave Ofset in ECGs”. Electron. Lett., 38 (14), pp. 682–683.
    https://doi.org/10.1049/el:20020499
15. Roslyakova, A.V., Chuprakov, P.G., 2012. “Sravnitel’nyy analiz algoritmov obnaruzheniya R-zubtsa elektrokardiogrammy”. Vyatskiy med. vestn., 2, pp. 29–35. (In Russian).
16. Faynzil’berg, L.S., 2013. “Kompyuternaya diagnostika po fazovomu portretu elektrokardiogrammy”. K.: Osvita Ukrainy, 191 p. (In Russian).
17. Clifford, G.D., 2006. “A Novel Framework for Signal Representation and Source Separation: Applications to Filtering and Segmentation of Biosignals”. J. of Biol. Syst., 14, pp. 169–183.
    https://doi.org/10.1142/S0218339006001830

Received 09.04.2015