Depth model for arrhythmia classification, and method and device utilizing model

Abstract

The invention provides a depth model for arrhythmia classification and a method and device using the model, and the depth model comprises a representation learning part and a sequence learning part. The representation learning part is used for receiving an equal-length sequence analyzed by the original electrocardiosignal; the representation learning part is constructed based on an MSCNN structure and is composed of two convolution block branches stacked in different scales; the convolution kernel of the first branch is large in scale and used for capturing low-frequency information of electrocardiosignals and outputting the low-frequency information in a multi-scale feature mode; the convolution kernel of the second branch is small in scale and used for capturing high-frequency information of the electrocardiosignal and outputting the high-frequency information in a multi-scale feature mode; the multi-scale feature output by the first branch and the multi-scale feature output by the second branch are spliced to form a multi-scale depth feature which is input to a sequence learning part; the sequence learning part is constructed on the basis of a Seq-Seq network taking LSTM as a basic unit, and an attention mechanism layer is arranged between an encoder and a decoder of the Seq-Seq network; the output is a time sequence depth feature.

Description

technical field [0001] The invention relates to arrhythmia detection technology, in particular to an automatic arrhythmia detection method without filtering noise reduction, abnormal value detection and manual feature extraction. Background technique [0002] According to statistics, death from acute heart disease accounts for nearly half of the total number of deaths from cardiovascular diseases, and nearly 31% of deaths worldwide are related to cardiovascular diseases. The main cause of sudden cardiac death is arrhythmia. Electrocardiography is the standard non-invasive tool for recording cardiac activity and is currently the most widely used and most reliable means of detecting arrhythmias. The arrhythmia classification of heartbeat signals not only takes a lot of time for cardiologists, but also increases the workload, which is still a relatively challenging task. This requires a lightweight automatic arrhythmia detection algorithm to provide auxiliary decision support ...

AI Technical Summary

Problems solved by technology

The algorithm based on template matching aims to use some abnormal heartbeats manually marked by doctors to automatically match other heartbeats with high similarity, which has the following defects: the algorithm is particularly sensitive to noise due to too simple matching operation; Can not give the doctor a complete preliminary result, the workload of the doctor is huge

The rule-based algorithm aims to use the medical knowledge related to electrocardiogram to automatically analyze the heartbeat, which has the following defects: the rules themselves are made by humans, and human beings may not have comprehensive knowledge of the electrocardiogram generated by a certain device; another The uncertainty of the electrocardiogram itself leads to the weak universality of the parameters in the rules

Therefore, these deep models tend to overfit, and often perform poorly on new test sets; 2) The high-precision classification results shown in most patents are only applicable to a single or even public training data set with a small amount of data, and there is a lack of The risk of robustness to heterogeneous data; 3) There are fewer categories of arrhythmia monitoring, and the standards of the American Association of Medical Instruments stipulate five commonly used arrhythmia types, but individual models are limited to the detection of specific heartbeat types; 4) the model The training overhead is huge, the calculation consumes a lot of resources, and there is still room for optimization

Images