Electroencephalogram signal classification method based on multi-channel feedback capsule network

Abstract

The invention discloses an electroencephalogram signal classification method based on a multi-channel feedback capsule network. The electroencephalogram signal classification method comprises the following steps: 1, carrying out data selection and slicing preprocessing on original electroencephalogram data; 2, establishing a multi-channel feedback capsule network classification model; 3, designing a loss function, and establishing a classification model optimization target; and 4, inputting data to train the network, and using the trained optimal model to complete electroencephalogram signal classification. The advantages of a feedback network and a capsule network are combined, signal classification can be automatically completed without manual feature extraction or signal processing of original electroencephalogram signals, the electroencephalogram signal classification accuracy can be remarkably improved, and therefore the application value of the electroencephalogram signals in the field of medical treatment and the like is increased.

Description

technical field [0001] The invention relates to the field of electroencephalogram signal classification, in particular to a method for automatically classifying and predicting raw electroencephalogram data of subjects through deep learning methods. Background technique [0002] The brain is an integral part of people's daily life, and the electrical activity in the cerebral cortex contains a wealth of information, which may contain information about different human emotions, motor imagination, and diseases. With the development of the field of brain-computer interface and intelligent medical treatment, EEG signals have been widely used in various fields such as emotional computing, motor imagery, and medical health. If we can fully tap the information of EEG information and accurately classify different EEG signals, the value of EEG signals in medical and other fields can be increased. [0003] An electroencephalogram (EEG) is a portable device that records the electrical a...

AI Technical Summary

Problems solved by technology

Although CNNs with different structures show different advantages in classification, it is difficult for CNN to describe the connection between local features, and the pooling operation will make it lose more spatial information, which is useful for multi-channel EEG signal classification tasks. say it might be crucial

The feature preprocessing process usually converts the original EEG data into various forms of features, and also includes operations such as filtering and denoising. Although cleaner data can be obtained, some important information may be lost at the same time.

Images