Training method and device of sleep stage identification model, terminal and medium

Abstract

The invention discloses a training method and device of a sleep stage identification model, a terminal and a medium. The training method comprises the following steps of acquiring a plurality of groups of electroencephalogram signals; performing frequency spectrum conversion on the plurality of groups of electroencephalogram signals to obtain a plurality of spectrograms; determining brain wave feature value sets corresponding to the plurality of spectrograms respectively; determining sleep stages respectively corresponding to the plurality of groups of brain wave feature value sets; and training a preset neural network model according to the plurality of groups of brain wave feature value sets and the respective sleep stages of the plurality of groups of brain wave feature value sets to obtain a sleep stage identification model. According to the training method and device of the sleep stage identification model, the terminal and the medium, a sample for training the neural network model has various characteristics of the electroencephalogram signals in a mode of calculating the electroencephalogram characteristic value set, so that the sleep stage identification model obtained by training can recognize the various characteristics of the electroencephalogram signals, the sleep stage identification accuracy is improved, and a data basis is provided for subsequently intervening a user in falling asleep.

Description

technical field [0001] The present application relates to the field of computer technology, in particular to a training method, device, terminal and medium for a sleep stage recognition model. Background technique [0002] Sleep stages are artificially delineated according to brain waves and physiological performance for the convenience of research. The international general method is to divide sleep into two different phases according to EEG performance, eye movement and muscle tension changes during sleep, namely non-rapid eye movement sleep NREM and rapid eye movement sleep REM. In related technologies, the following methods are mainly used to identify sleep stages: 1. Obtain a spectrogram of the EEG signal first, and then professionals make judgments based on the spectrogram; 2. Judgment is made through a sleep analysis model. Of these two methods, the former has the problem of high professional requirements for users and is generally used in clinical practice; the latt...

AI Technical Summary

Problems solved by technology

Of these two methods, the former has the problem of high professional requirements for users and is generally used in clinical practice; the latter is generally obtained through training on a certain feature of the EEG signal, and has the problems of low recognition accuracy and poor accuracy.

Images