Mi EEG signal recognition method based on feature fusion and particle swarm optimization algorithm

Abstract

The invention discloses a MI EEG signal recognition method based on feature fusion and particle swarm optimization algorithm, comprising the following steps: S1, collecting MI EEG signals, performing band-pass filtering on the collected MI EEG signals, and then passing The wavelet soft threshold method is used for denoising operation, and the EEG characteristic signal is extracted; S2, PSO‑RF is used to perform feature screening on the EEG characteristic signal. The present invention combines band-pass filtering, wavelet denoising, channel screening, feature extraction, feature fusion, feature selection, and pattern classification, effectively integrates these seven parts, and finally obtains an integrated classifier that can achieve an average correctness of 98.34%. rate, and the AUC value and F-score are also excellent, so it can achieve the purpose of accurate motor imagery classification.

Description

technical field [0001] The invention relates to an MI electroencephalogram signal recognition method based on feature fusion and particle swarm optimization algorithm. Background technique [0002] Brain Computer Interface (BCI), as an emerging technology, is gradually playing its role in military, entertainment, and medical rehabilitation. As an important application of BCI, motor imagery (MI) mainly collects the EEG signals of subjects imagining limb movements, and uses machine learning (Machine Learning, ML) to classify, and finally feeds back the classification results to External equipment assists subjects to perform limb movements and helps physically disabled people perform daily movements. Therefore, this research direction has great significance in the field of medical rehabilitation. [0003] Among the numerous brain-computer interaction control paradigms, motor imagery-based brain-computer interface is one of the most common types. Motor imagery is to generate c...

AI Technical Summary

Problems solved by technology

[0004] Due to the low signal-to-noise ratio of the EEG signal and the lack of feature information obtained by a single feature extraction method, it is impossible to improve the classification accuracy of the brain-computer interface.

Images