Epileptic state recognition method based on transfer learning and cavity convolution

Abstract

The invention provides an epileptic state recognition method based on transfer learning and cavity convolution. The epileptic state recognition method comprises the following steps: S1, extracting a plurality of wavelet packet coefficient groups of each group of original epileptic electroencephalogram signals under a specific frequency as a feature group; s2, removing significantly related waveletpacket coefficient groups in the feature groups to realize dimensionality reduction of the feature groups, wherein each wavelet packet coefficient of the feature group after dimension reduction is afeature value; s3, standardizing all characteristic values extracted from the plurality of groups of original epilepsy electroencephalogram signals; s4, taking all the characteristic values subjectedto standardization processing as a test data set, and taking characteristics in an existing epilepsy electroencephalogram signal characteristic database as a training data set; achieving cross-domainknowledge migration through an improved CMJAE migration learning method, taking a two-dimensional hole convolutional neural network as a classifier, and iteratively acquiring a classification result of the test data set; and S5, verifying the classification accuracy by adopting a ten-fold cross validation method.

Description

technical field [0001] The invention relates to feature extraction, pattern classification, migration learning and deep learning of non-stationary signals, and belongs to the technical field of signal processing and pattern recognition. Background technique [0002] Currently, epilepsy is a common neurological disease. According to the World Health Organization (WHO), nearly 2.4 million people worldwide are diagnosed with epilepsy every year. trend. Epilepsy is mainly a brain dysfunction caused by abnormal synchronous discharge activities of a large number of neuron clusters. Scalp EEG or intracranial EEG reflects the main brain nerve activities and contains a large amount of physiological and pathological information in the brain. Therefore, the intelligent recognition of EEG (electroencephalogram) signals is the main means of studying epileptic seizures. [0003] At present, the diagnosis of epilepsy requires the manual analysis of EEG signals of patients for several day...

AI Technical Summary

Problems solved by technology

[0008] Aiming at the problem of insufficient training data and low classification accuracy caused by inconsistent distribution of training data and test data in epilepsy EEG signal recognition, the present invention proposes an epilepsy EEG signal recognition method

Images