There have been several methods for determining T wave-end. But none of them can overcome the difficulty from multiformity of ECG signal pattern. In this paper, a method for determining T wave-end using evolutionary algorithm (EA) is proposed. In this way, first, every characteristic parameter related to T wave-end is encoded to a string of codes, and adaptation function is constructed with the string of codes. Then choose the individual according to the adaptation function value and do genetic operation (reproduction, crossover and mutation), so as to produce offspring with more adaptation function value. Because of EA's autoadaptation and autoorganization character, it can trace ECG signal type and find the T wave-end automatically.Experiment results show that the error ratio of recognizing T wave-end using this method is much smaller than that using existing method.
Luo Jianhua Li Gaoping Zhuang Tiange(Department of Biomedical Engineering, Shanghai Jiaotong University, Shanghai 200030) (ECG Department of Zhongshan Hospital, Shanghai Medical University, Shanghai 200030)
