Title：Sleep stage recognition method and system based on adversarial learning
Country：China
Patent No.：202010408610.0
Legal Status：Authorized
Inventor：Milin Zhang, Yiqiao Liao, Xiang Xie, Zhihua Wang
Assignee：Tsinghua University
Address：Tsinghua University,Haidian District Beijing 100084, China
Filing Date：2020-05-14
Issue Date：2021-07-20
Abstract：
The invention discloses a sleep stage recognition method and system based on adversarial learning. The method comprises the following steps: inputting an EEG signal training set to a to-be-trained adversarial feature extractor, and carrying out training updating based on adversarial learning to obtain an adversarial feature extractor; inputting the EEG signal training set into a to-be-trained individual feature extractor for training and updating to obtain an individual feature extractor; respectively inputting the EEG signal test set into an adversarial feature extractor and an individual feature extractor, and training and updating a to-be-trained neural network, a to-be-trained sleep stage classifier and a to-be-trained conditional random field layer to obtain a neural network, a sleepstage classifier and a conditional random field layer; and inputting the EEG signal to be classified into a three-feature network, and determining a sleep stage. The equipment size can be reduced by only using a single-channel EEG signal; the universality of the extracted features is improved through adversarial learning, the difference of EEG data of different subjects is reduced, and the accuracy is improved; and professional knowledge is not needed, and the operation is simple.
Patent Certificate：
PDF/Jpg