Electrocardiosignal anomaly detection method, system and device and storage medium

Abstract

The invention discloses an electrocardiosignal anomaly detection method, system and device and a storage medium. The electrocardiosignal anomaly detection method comprises the steps of obtaining first data; preprocessing the first data to generate second data; obtaining a preset number of samples from the second data as first samples; inputting the first sample into the grid model for training, and generating an encoder; performing feature processing and dimension reduction on the first data through an encoder to generate a second sample; and according to the second sample, generating an electrocardiosignal detection result through a hidden Markov model. According to the method, the encoder is obtained by training the grid model integrated by the bidirectional circulation gating unit layer and the attention layer, the data processing capacity of the encoder is improved, and the encoder can extract more detail features from original electrocardiosignal data; the hidden Markov model with the Watson hybrid model as the emission density is adopted to recognize the sample, and the accuracy and comprehensiveness of the electrocardiosignal detection result are improved.

Description

technical field [0001] The present application relates to the technical field of detection, in particular to a method, system, device and storage medium for abnormal detection of ECG signals. Background technique [0002] The traditional detection of ECG abnormality usually requires doctors with rich clinical experience to read the images manually with the naked eye. The degree of automation is low, and there are problems such as low detection efficiency and misdiagnosis due to doctor fatigue. Most of the existing ECG automatic detection methods use neural network models. The neural network model is a supervised classification method, which requires a lot of manpower to mark a large number of labeled ECG data, and the accuracy is not high in the case of small sample data. [0003] At present, some ECG signal abnormality detection methods use unsupervised deep models or data modeling. [0004] Among them, the network layer of the self-encoding hidden layer of the unsupervis...

AI Technical Summary

Problems solved by technology

This method is a supervised learning process. In the stage of training model parameters, a large amount of labeled data needs to be used, which consumes a lot of manpower and material resources.

Moreover, the annotation process of data labels is highly specialized and takes a lot of time, which makes the generalization of the model not high; secondly, in the classification process, the model needs to calculate the distance between the feature vector and the center of each cluster every time. It is very time-consuming for a large number of long-term series data sets, resulting in slow model convergence, which is not conducive to the application in portable ECG equipment

Images