Cardiovascular disease and its consequences on human health have never stopped and even show a trend of appearing in increasingly younger generations. The establishment of an excellent deep learning algorithm model to assist physicians in identifying and the early screening of ECG abnormalities can effectively improve the accuracy of diagnosis. Therefore, in this study, the deep residual network model is adapted for feature extraction and classification of ECG signals by pooling embedded into layers and double channel connection. At the same time, the wavelet adaptive threshold denoising algorithm is used to complete the high signal-to-noise filtering of ECG signals. Then, the alternate pooling residual network (APRN) is compared with the convolutional neural network (CNN), CNN with one residual unit (CNN-R), and the deep residual network (ResNet-18) using ECG datasets from the American MIT-BIH arrhythmia and ST segment abnormality database, European ST-T database, and sudden cardiac death ambulatory ECG database. The results are as follows: The average classification accuracy of the APRN on the four datasets is 97.89%, while the accuracies on CNN, CNN-R, and ResNet-18 are 97.17%, 97.53%, and 97.73%, respectively. In addition, compared with ResNet-18, the classification accuracy of our APRN on each class of data improves by 16.44% in total.

Cardiovascular disease has always been one of the major threats to human health. The recognition of ECG and PCG signals, which are physiological signals generated when the heart beats, can greatly improve the efficiency of cardiovascular disease detection using deep learning techniques. Based on this, it is proposed that a simple and effective pooling convolution model for multi-classification of ECG and PCG signals. Firstly, ECG, PCG and synchronized ECG-PCG data are pre-processed. Then, several structure blocks are designed, including the block stacked by max-pooling layer(MCM), convolution layer and max-pooling layer, its multiple variants and the residual block(REC). All network models that use these structure blocks separately are based on the ResNet-18 framework. By changing the number of structure blocks, the model can be applied to ECG and PCG data at different sampling rates. The final test shows that the accuracy of the model using the MCM structure block is 98.70%, 92.58% and 99.19% respectively on the ECG, PCG fusion dataset and synchronized ECG-PCG dataset, which is higher than all the networks using its multiple variants. At the same time, the accuracy is improved by 0.02%, 4.30% and 1.43% compared to the model using the REC structure block. In addition, this work also carries out tests on multiple ECG and PCG datasets and compares them with other published literature, further verifying that the model using the MCM structure block has a higher detection rate for ECG and PCG signals.