Heart beat screening method

The embodiment of the invention discloses a cardiac beat screening method, which comprises the following steps: pre-training a deep recursive residual network model for screening out an optimal cardiac beat in electrocardiosignals according to input electrocardiosignals, then obtaining to-be-screened electrocardiosignals, and inputting the to-be-screened electrocardiosignals into the preset deep recursive residual network model; according to the method, the cardiac beat waveform which is not influenced by factors such as interference or arrhythmia or the cardiac beat waveform which is slightly influenced by the factors such as interference or arrhythmia can be quickly screened out, so that when the optimal cardiac beat is used for ST analysis, the change of the calculated final ST value is smaller; therefore, the stability and the accuracy of the calculated final ST value can be ensured. 本发明实施例公开了一种心拍筛选方法，包括：通过预先训练好一个用于根据输入的心电信号，筛选出心电信号中最优的心拍的深度递归残差网络模型，然后获取待筛选心电信号，并将待筛选心电信号输入至预设深度递归残差网络模型中，即可得到最优心拍，该方式可以快速的筛