Sleep stage staging system combining with rapid representation learning and semantic learning

Abstract

The invention provides a sleep stage staging system combining with rapid representation learning and semantic learning. The staging system includes a signal acquisition unit, a data pre-processing unit, a signal processing unit and a prediction result processing unit; the signal acquisition unit includes an EEG signal acquisition module and a knowledge data acquisition module which are respectively used for obtaining original EEG signals and relevant knowledge of sleep stage staging; the data pre-processing unit includes a feature preprocessing module and a knowledge pre-processing module which are respectively used for pre-processing on the collected EEG signals and knowledge data; the signal processing unit includes a representation learning module and a semantic learning module, so thatEEG signal features and semantic features related to sleep staging can be effectively extracted, and two classification results can be obtained; and the prediction result processing unit weights andfuses the obtained two classification results so as to obtain a final result. The system accelerates training and prediction speed and effectively utilizes the signal features and semantic features, so that the accuracy of sleep stage staging can be greatly enhanced.

Description

technical field [0001] The invention relates to the fields of signal processing and sleep stage staging, in particular to a sleep stage staging system combining fast representation learning and semantic learning. Background technique [0002] Traditional sleep stage scoring based on doctor's observation is very tedious, time-consuming and subjective, which requires the doctor to analyze the signal in the PSG recording to derive a sleep score of about 8 hours. Therefore, many automatic sleep assessment methods have been proposed. These studies first extract various features related to sleep from EEG signals, such as time domain features, frequency features, correlation features, entropy features, etc., and then use machine learning methods (decision tree, support vector machine, etc.) to analyze the extracted features. sort. However, these feature extraction methods are subjective, and many potential sleep features have not been mined, and these latent sleep features can pl...

AI Technical Summary

Problems solved by technology

However, there is no way to effectively use this knowledge to improve the classification accuracy.

Images