Electrocardiosignal noise reduction method based on adversarial generative network

According to the electrocardiosignal noise reduction method based on the generative adversarial network, the conditional generative adversarial network is used, so that the directionality of generateddata and the matching performance of a signal after noise reduction and a noisy signal are ensured, and the generalization ability of a model is improved. In the training process, uncertain noise mixed data is used for training, so that the model can perform noise reduction on various noise mixed signals without pre-judgment, and the complexity of the method is simplified. Design of an improved loss function is realized, the resistance in the original loss function of the CGAN is reserved, the root-mean-square error and the noise-to-signal power ratio are increased, the increase of the root-mean-square error enables the model to capture the local features of the signal and maintain the useful medical features of the signal, and the increase of the noise-to-signal power ratio enables the model to capture the globa