CLASSIFICATION FOR UTERINE EMG SIGNALS: COMPARISON BETWEEN AR MODEL AND STATISTICAL CLASSIFICATION METHOD

Abstract —This article proposes a method for modeling andclassification apply on the uterine contractions in the electromyo-gram (EMG) signal for the detection of preterm birth. Thefrequency content of the contraction changes from one woman toanother and during pregnancy. Firstly we apply an AR model onthe Uterine EMG signal for the calculation of the a i parameters.Wavelet decomposition is used to extract the parameters of eachsimulated contraction, and an unsupervised statistical classifica-tion method based on Fisher test is used to classify the signals. Aprincipal component analysis projection is then used to evidencethe groups resulting from this classification. Results show thatuterine contractions may be classified into independent groupsaccording to their frequency content and according to term (atthe recording, or at delivery). Copyright ° c 2007 Yang’s ScientificResearch Institute, LLC. All rights reserved.Index Terms —AR model, wavelet, uterine EMG, classification,preterm birth. I. I

[2]  J. P. Morucci,et al.  External recording and processing of fast electrical activity of the uterus in human parturition , 1984, Medical and Biological Engineering and Computing.

[3]  R. M. Maiden,et al.  The electrical potentials of the human uterus in labor. , 1946, American journal of obstetrics and gynecology.

[4]  C. Li,et al.  Detection of ECG characteristic points using wavelet transforms. , 1995, IEEE transactions on bio-medical engineering.

[5]  Steven Kay,et al.  Modern Spectral Estimation: Theory and Application , 1988 .

[6]  C. Marque,et al.  Use of the electrohysterogram signal for characterization of contractions during pregnancy , 1999, IEEE Transactions on Biomedical Engineering.

[7]  G. Wolfs,et al.  Electromyographic Observations on the Human Uterus during Labour , 1979, Acta obstetricia et gynecologica Scandinavica. Supplement.

[8]  Dominique Devedeux.  Évaluation quantitative de certaines caractéristiques de distributions temps/fréquence : application à l'EMG utérin , 1995 .

[9]  Stéphane Bounan,et al.  Menace d'accouchement prématuré , 2006 .

[10]  M Katz,et al.  Anteparturn Arnbulatorv Tocodvnamometry: The Significance of Low‐Amplitude, High‐Frequency Contractions , 1987, Obstetrics and gynecology.

[11]  Laurent Miclet.  Reconnaissance des formes , 1984 .

[12]  G J HERTSCH,et al.  Electrical activity of the human uterus in labor; the electrohysterograph. , 1950, American Journal of Obstetrics and Gynecology.

[13]  J. Alexandre,et al.  222 – Menace d'accouchement prématuré , 2009 .

[14]  Jacques Duchêne,et al.  Detection and classification of multiple events in piecewise stationary signals: Comparison between autoregressive and multiscale approaches , 1999, Signal Process..

[15]  M. Khalil,et al.  An unsupervised classification method of uterine electromyography signals using wavelet decomposition , 2004, The 26th Annual International Conference of the IEEE Engineering in Medicine and Biology Society.

[16]  C. Marque,et al.  Unsupervised Classification in Uterine Electromyography Signal: Toward The Detection of Preterm Birth , 2005, 2005 IEEE Engineering in Medicine and Biology 27th Annual Conference.

[17]  Andreas Baudisch.  Classification and Interpretation , 1989, J. Symb. Log..

[18]  E. Hon,et al.  Cutaneous and uterine electrical potentials in labor; an experiment. , 1958, Obstetrics and Gynecology.

[19]  Jacques Duchêne,et al.  Uterine EMG analysis: a dynamic approach for change detection and classification , 2000, IEEE Transactions on Biomedical Engineering.

[20]  Mohamad Khalil.  Une approche de la détection et de la classification dans les signaux non stationnaires : application à l'EGM uterin , 1999 .

[21]  Catherine Marque,et al.  Uterine EHG Processing for Obstetrical Monitorng , 1986, IEEE Transactions on Biomedical Engineering.