Análise comparativa de desempenho das transformadas Wavelet e Hilbert na detecção do QRS em ECG

João Paulo do Vale Madeiro, Paulo César Cortez, João Alexandre Lôbo Marques
2009 Revista Brasileira de Engenharia Biomédica  
Resumo O processo de detecção do complexo QRS é o primeiro passo de um processo de extração de parâmetros do sinal eletrocardiograma (ECG) em sistemas de auxílio ao diagnóstico médico. O presente trabalho apresenta resultados detalhados de comparação da aplicação de duas transformadas matemáticas, Wavelet e Hilbert, em um algoritmo de detecção de QRS em termos de taxas de detecções corretas (sensibilidade e preditividade positiva) e de uma medida de frequência de recorrência a processos de
more » ... agem (pré-processamento). Uma abordagem inovadora é implementada, na qual as rotinas de filtragem são inseridas dentro do estágio de decisão, ou seja, é realizada a supressão da etapa de préprocessamento. As transformadas são aplicadas no algoritmo, que é baseado em um limiar adaptativo, com o objetivo de realçar, apenas quando necessário, os picos (pontos fiduciais) do QRS. Em uma primeira abordagem, apenas a transformada Wavelet é utilizada neste realce e, numa se gunda abordagem, a transformada de Hilbert é inserida em série à aplicação da Wavelet em dois possíveis arranjos. São realizados experimentos dos algoritmos sobre os exames da base de dados Arrhythmia Database, pertencente ao conjunto de bases de dados do MIT-BIH. É composta por 48 gravações de ECG com duração de trinta minutos, amostrados a uma frequência de 360 Hz com resolução de 4,88 μV sobre uma faixa de variação de 10 mV. Ao todo, contabilizam-se 109.662 complexos QRS. Taxas de 98,85% de sensibilidade e 95,10% de preditividade positiva são obtidas com a aplicação exclusiva da transformada Wavelet, enquanto que 98,89% de sensibilidade e 98,52% de preditividade positiva são obtidas com a aplicação em série das transformadas Wavelet e de Hilbert. Palavras-chave: Eletrocardiograma (ECG), Complexo QRS, Transformada Wavelet (TW), Transformada de Hilbert (TH). Abstract The process of QRS detection is the first stage of a greater process: the feature extraction in the electrocardiogram (ECG). This work presents detailed results on the performance of two mathematical transforms, Hilbert and Wavelet, which are applied in QRS detection. The evaluation parameters are the detection rates and a measure of frequency of recurrence to filtering processes. An innovative approach is implemented: the filtering routines are inserted in the decision stage, i.e. the preprocessing stage is removed. The algorithm is based on adaptive threshold technique and the two transforms are applied in order to emphasize, only when necessary, the QRS fiducial points. In a first approach, only the Wavelet transform is applied, and in a second approach, the Hilbert transform is inserted before the Wavelet transform or after it. We evaluate these approaches on the well-known MIT-BIH Arrhythmia Database. It contains 48 half-hour recordings of annotated ECG with a sampling rate of 360 Hz and 4.88 μV resolution over a 10 mV range, totalizing 109,662 QRS complexes. Sensitivity rates of 98.85% and 98.89% are respectively attained when the Wavelet transform is applied in the filtering processes and both Hilbert and Wavelet transforms are applied. Predictability rates of 95.10% and 98.52% are also attained respectively using Wavelet transform and the simultaneous application of Hilbert and Wavelet transforms in the filtering processes. Análise comparativa de desempenho das transformadas Wavelet e Hilbert na detecção do QRS em ECG Performance comparison analysis of Wavelet and Hilbert transforms for QRS detection in ECG http://dx.doi.org/10.4322/rbeb.2012.072 Desempenho das transformadas Wavelet e Hilbert na detecção do QRS em ECG Madeiro, J. P. V.; Cortez, P. C.; Marques, J. A. L. Rev. Bras. Eng. Biom., v. 25, n. 3, p. 153-166, dez. 2009 Braz. J. Biom. Eng., 25(3): 153-166, Dec. 2009 Extended Abstract Introduction A generalized scheme for QRS detection in ECG waveforms, developed by the first systems and now shared by many algorithms, is divided into a preprocessing or feature extraction stage, including linear and nonlinear filtering, and a decision stage including peak detection and decision logic (Kohler et al., 2002) . As examples of linear transformations, we emphasize Wavelet and Hilbert transforms, which are used at this work. The Wavelet transform, which is explored since the 90's, remains as the favorite tool of time-frequency analysis (Addison, 2005) . Through the application of DWT (Discrete Wavelet Transform), different approaches search for critical points, e.g. "maximum modulus lines" exceeding some thresholds at different scales (Martínez et al., 2004) . Hilbert transform, also used since the 90's, is applied on the calculus of signal envelope and it is usually preceded by a derivative filter (Arzeno et al., 2008; Benitez et al., 2001) (Arzeno et al., 2008) . . The effects of the Hilbert transform have been explained in terms of its odd symmetry property and signal envelope. The odd-phase component of the filter provides a rectification of the differentiated ECG signal in order to identify the QRS peaks while the uniform magnitude of the filter ensures that necessary information of the QRS complexes is preserved An innovative methodology is proposed in this work where we seek to abolish the preprocessing stage, pointed as an essential prerequisite by the literature of software QRS detection algorithms. Instead of filtering the entire length of the signal, the application of Hilbert and Wavelet transforms is inserted into the decision stage. Free of any preprocessing, the ECG signal is the input information of the decision stage, except for an initial interval, lasting about 10 seconds. This initial interval is preprocessed in the training stage, which applies Hilbert and Wavelet transforms, as illustrated in the block diagram of Figure 1 . The analysis stage is based on an adaptive threshold and a previous knowledge of QRS complex statistics, which is provided by the training stage: average amplitude of the QRS fiducial points and average and standard deviation of intervals between beats. We only apply filtering, through Hilbert and Wavelet transforms, in certain regions of the signal, which are identified by a subsequent decision routine, in order to correct possible failures of the adaptive threshold: false-positive or false-negative. In this paper, two approaches for QRS detection are applied over the ECG signal, without the implementation of filtering on the complete signal, but only in certain segments, which are identified by the algorithms. The first approach uses the Wavelet transform, which is applied simultaneously with the adaptive threshold technique. The second approach performs consecutive applications of Wavelet and Hilbert transforms in two possible arrangements and they are also applied simultaneously with the adaptive threshold technique. Performances of the two approaches are compared through experimental tests over Arrhythmia database (MIT-BIH, 2009). Desempenho das transformadas Wavelet e Hilbert na detecção do QRS em ECG Madeiro, J. P. V.; Cortez, P. C.; Marques, J. A. L. Rev. Bras. Eng. Biom., v. 25, n. 3, p. 153-166, dez. 2009 Braz. J. Biom. Eng., 25(3): 153-166, Dec. 2009 Desempenho das transformadas Wavelet e Hilbert na detecção do QRS em ECG Madeiro, J. P. V.; Cortez, P. C.; Marques, J. A. L. Rev. Bras. Eng. Biom., v. 25, n. 3, p. 153-166, dez. 2009 Braz. J. Biom. Eng., 25(3): 153-166, Dec. 2009 Desempenho das transformadas Wavelet e Hilbert na detecção do QRS em ECG Madeiro, J. P. V.; Cortez, P. C.; Marques, J. A. L. Rev. Bras. Eng. Biom., v. 25, n. 3, p. 153-166, dez. 2009 Braz.
doi:10.4322/rbeb.2012.072 fatcat:esqkhij5ufhfpbbn3mb5r6grqi