Abstract

Info

Author(s):: ZHANG Xuan; LIU Ankang; ZHANG Peiling; School of Physics and Electronic Information, Henan Polytechnic University, Jiaozuo 454000, China

Keywords:: Keywords: epilepsy electroencephalogram signal embedded artificial intelligence wavelet packet decomposition one-dimensional convolutional neural network WeChat applet

Abstract:: Abstract: Epilepsy monitoring aims to prevent accidents that patients may experience due to unconsciousness during epileptic seizures. Epilepsy can be monitored in real time by analyzing EEG signals, so as to provide corresponding references for the diagnosis, treatment and evaluation of epilepsy.A seizure monitoring system based on embedded AI is designed in the study, and the system is divided into 3 modules, namely training module, test module and alarm module. Born data set is adopted in training module, and wavelet packet decomposition and one-dimensional convolutional neural network are used for training. The final accuracy of the model is up to 98.3%. The test module uses brainwave sensor to collect signals which are transmitted through Bluetooth, and then compared with the training model after the test model is processed by MCU. The alarm module will feedback the above results to the WeChat applet, and alarm in time if there is any abnormality. The system designed based on embedded AI can reduce the injury to patients and protect the safety of patients for it adopts wearable epilepsy monitoring and alarm equipment and is capable of real-time monitoring of epileptic seizures.