Cardiac arrythmias classification algorithm based on convolutional neural network

Abstract

The invention discloses a cardiac arrythmias classification algorithm based on a convolutional neural network. The cardiac arrythmias classification algorithm comprises a small-scale Deep-LeNet network suitable for a sparse image of electrocardiosignals. According to the characteristics of a small convolutional core, the time consumed for network classification is shorter, and the accuracy is improved. Then, the convolutional neural network is put forward. The width of the network can be increased, the adaptability of the network to the scale is increased, the network is more suitable for recognizing the sparse image. Under the condition of quite slightly increasing the network time consumption, the classification accuracy of the network can be greatly increased; due to the integrated recognition and classification process, the cardiac arrythmias classification algorithm can be better used for family-practice-level diagnosis and has great significance in accurate recognition of cardiacarrythmias.

Description

technical field [0001] The invention belongs to the field of biomedical signal recognition, and relates to a ECG signal image classification technology, in particular to an ECG signal image feature extraction and classification technology based on a convolutional neural network, and specifically relates to building an ECG signal image classification technology. Convolutional neural network. Background technique [0002] The ECG contains a wealth of pathological information about cardiac activity, the most important analysis of which is the classification of heart beats, which is very important for detecting arrhythmias. Since it is very time-consuming and impractical to diagnose arrhythmia by artificially analyzing long-term ECG signals, it is worth studying to use automatic algorithms to assist in the diagnosis of arrhythmia. [0003] Convolutional Neural Networks (CNN) are a new form of deep learning in which the network structure consists of many hidden layers and parame...

AI Technical Summary

Problems solved by technology

[0007] In view of the current situation and existing problems in the identification of the above-mentioned existing arrhythmia, the purpose of the present invention is to propose a CNN network suitable for ECG signal classification, which can not only reduce the overall time-consuming of the network, but also obtain higher accuracy. rate, to overcome the complex network and time-consuming problems existing in the current ECG signal classification technology

Images