Research On The Key Technologies Of T Wave Alternans Detection In ECG

Cardiovascular diseases have become the top killer of human life, with an awfully high fatality. There are nearly600000cases of Sudden-Cardiac Death (SCD) in China every year. Hence, how to predict SCD in advance has become the focus of the medical circle at present. A large number of clinic medical research demonstrate that T-wave alternans(TWA) in electrocardiograph is always accompanied by such hear disorder as SCD, arrhythmia, etc, thus TWA has become the non-invasive indicator predicting the two diseases aforesaid. Therefore, the research on key technology of TWA detection is of great clinic value and social significance.The paper has made the key technologies of T-wave alternans as the main target, which may foucus on the research of pre-testing and alternating phase of algorithm.The pre-processing stage:as the traditional R-wave identification method which is vnlnerable to noise and a significant T-wave effect may arouse the problem of missed and false detection, this paper puts forward an adptive threshold algorithym which based on difference. It improves the recognitive accuracy of T-wave through the RR interval threshold, differential threshold and R-square wave amplitude R-wave threshold from the aspacts of the signal itself, the differential signal and the signal difference of the square three dimensions. Finally, it verifies the validity and accuracy of the algorithm by the use of simulation experiments.TWA detection stage:the conventional TWA detection time-domain algorithm is either sensitive to power frequency disturbance and baseline drift or T-wave alignment, while the frequency domain method has no temporal resolution despite lower demand on signal quality, unable to trace unstable TWA. In view of this, this paper proposes TWA alternate detection algorithm based on Kalman optimal estimate, which deals with the T-wave grouped by parity and applies dynamic tracing strategy to detect TWA taking advantage of the good performance of Kalman filter in inhibiting Gaussian noise and accurate tracing. The simulation experiment demonstrates that this algorithm can detect alternation accurately as well as inhabit high frequency noise and baseline drift in an effective way and identify unsteady TWA.Due to the unsteadiness of TWA, ECG signals include colored noise in addition to Gaussian noise and. In view of the above, this paper further raises the steady T-wave alternate detection algorithm based on wavelet transform and Boostrap method. This algorithm first uses wavelet transform to effectively divide unsteady TWA into several segments of steady TWA, then employs Bootstrap method to make non-parameter estimate of these segments to detect TWA. It is demonstrated by simulation experiment that this algorithm is advantageous in TWA detection of unsteady TWA including colored noise.