A single-channel EEG epileptic spike detection method and device based on U-Net network

Through a multi-task deep learning model based on U-Net network, automatic detection and positioning of electroencephalogram spike waves is realized, solving the problem of time-consuming manual review and difficult to accurately locate automatic detection in the prior art, and improving the accuracy and readability of detection.

CN118380139BActive Publication Date: 2025-03-18TIANJIN UNIV
3 Cites 0 Cited by

Patent Information

Application Number
CN202410512749.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-04-26
Publication Date
2025-03-18
Estimated Expiration
2044-04-26

AI Technical Summary

Technical Problem

The existing electroencephalogram spike wave detection methods rely on manual review to consume time and labor-intensive work, and automatic detection methods are difficult to accurately locate the spike wave position, resulting in missed detection.

Method used

Using a single-channel EEG epilepsy spike wave detection method based on U-Net network, by establishing a multi-task deep learning model, it can not only classify whether the EEG fragment contains spike waves, but also segment the location of the spike waves.

Benefits of technology

The accuracy and readability of the spike wave detection results are improved, and the defects of incomplete manual screening characteristics in traditional methods and the inability to locate the spike wave position in deep learning methods are overcome.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The present invention provides a single-channel EEG epilepsy spike detection method and device based on U‑Net network, which belongs to the field of intelligent medical technology. In the training stage, an EEG signal with a spike wave position marked is obtained as an input sample, and preprocessing is performed to obtain a preprocessed EEG sample. The preprocessed EEG sample and spike wave label information are packaged and input into a model training module, and a multi-task deep learning model based on U‑Net is constructed to optimize the training and save the optimal parameters; in the verification stage, an EEG signal with an unmarked spike wave position is used as an input sample, and preprocessing is performed. The preprocessed EEG sample is input into the trained model for detection, and the detection result is input into the post-processing module for further processing to obtain the corrected spike wave position. The present invention can not only improve the accuracy of the EEG spike wave detection algorithm, but also provide useful help for the visualization of the spike wave position.
Need to check novelty before this filing date? Find Prior Art