Electrocardiogram (ECG) signal feature extraction is important in diagnosing cardiovascular diseases. This paper presents a new method for nonlinear feature extraction of ECG signals by combining principal component analysis (PCA) and kernel independent component analysis (KICA). The proposed method first uses PCA to decrease the dimensions of the ECG signal training set and then employs KICA to cal-B Hongqiang Li Circuits Syst Signal Process culate the feature space for extracting the nonlinear features. Support vector machine (SVM) is utilized to determine the nonlinear features of the ECG signal classification. Genetic algorithm is also used to optimize the SVM parameters. The proposed method is advantageous because it does not require a huge amount of sampling data, and this technique is better than traditional strategies to select optimal features in the multi-domain feature space. Computer simulations reveal that the proposed method yields more satisfactory classification results on the MIT-BIH arrhythmia database, reaching an overall accuracy of 97.78 %.

Electrocardiogram (ECG) signal classification is an important diagnosis tool wherein feature extraction plays a crucial function. This paper proposes a novel method for the nonlinear feature extraction of ECG signals by combining wavelet packet decomposition (WPD) and approximate entropy (ApEn). The proposed method first uses WPD to decompose ECG signals into different frequency bands and then calculates the ApEn of each wavelet packet coefficient as a feature vector. A support vector machine (SVM) classifier is used for the classification. The particle swarm optimization algorithm is used to optimize the SVM parameters. The proposed method does not require dimensionality reduction, has fast calculation speed, and requires sim-B Hongqiang Li Circuits Syst Signal Process ple computations. The classification of the signals into five beats yields an acceptable accuracy of 97.78 %.

In most traditional electrocardiogram (ECG) detection procedures, wet electrodes must be glued to the skin during the procedure and may cause problems such as inconvenience and skin irritation. Furthermore, the quality of the acquired signals decreases because the glue dehydrates over time. In this study, a non-contact ECG acquisition system based on capacitive coupling textile electrodes with low-power consumption and high input impedance is presented. We designed electrodes that have a composite and textile structure. A kind of conductive textile with stainless steel wire creates these electrodes. We wove the conductive textile that has good electrical conductivity with a surface resistivity of 1.25 Ω/sq. Both circuit models of the skin–electrode interface and amplifier for the capacitively coupled textile electrode were established, and the output signal-to-noise ratio (SNR) of the front-end circuit was proposed. The integrated system combines amplification, filter circuit and analogue-to-digital converter. The final measurement shows that the ECG signals acquired by our system are adequate for heartbeat detection and applicable to clinical practice.

The electrocardiogram (ECG) is an important technique for heart disease diagnosis. This paper proposes a novel method for ECG beat classification. Several important issues exist in the ECG beat classification, which, if suitably addressed, may lead to development of more robust and efficient recognizers. Some of these issues include feature extraction, choice of classification approach and optimization. A new method for non-linear feature extraction of ECG signals based on empirical mode decomposition (EMD), approximate entropy (ApEn) and wavelet packet entropy is presented. The proposed method first uses EMD to decompose ECG signals into a finite number of intrinsic mode functions (IMFs), calculates the ApEn of IMFs as one feature and then obtains the wavelet packet entropy of wavelet packet coefficients as another feature. The two features are regarded as a feature vector. The support vector machine (SVM) and probabilistic neural network (PNN) are used for beat classification. The particle swarm optimization algorithm is used to optimize parameters of the PNN and SVM. The performance of the SVM classifier is slightly superior to that of the PNN classifier with 98.6% accuracy.