基于滑动窗口的两类运动想象EEG神经网络识别研究
首发时间:2018-05-18
摘要:本文采用信号加窗与神经网络相结合的方式对两类运动想象脑电信号进行识别,对相关问题展开分析研究,通过与传统的滑动平均方法做对比,分析这种方法对脑电识别效果的改善程度及特点,以及如何将这种方法更好地用于脑电识别。研究显示,滑动窗口的信号处理方式与神经网络方法的结合对脑电信号的识别是有利的,使识别效果在对脑电信号进行滑动平均及线性识别基础上进一步得到加强,也更趋于稳定。此外,该方法的识别效果受窗口宽度影响,适合的窗口设置能够更好地展现这种方法的优势与实用价值。
For information in English, please click here
Research on Neural Network Identification of Two-Type Motor Imagery EEG Based on Sliding Windows
Abstract:EEG of two-type motor imagery was identified by a kind of combining method of signal windowing and neural network in the paper, and some relevant questions were also researched. By comparing with the traditional moving average methods, some questions were analyzed including the improvement degree and characteristics of this method to EEG identification results, and how to use the method on EEG identification better. Studies showed that it was helpful to EEG identification to combine the signal processing style of sliding window and the method of neural network. Such combining method enhanced identification result compared with the moving average processing and the linear identification of EEG, and made identification result more stabilized. What\'s more, the identification result of this method was affected by window width, and suitable window setting can make this method exert its advantages and practical values better.
Keywords: BEI (Biotic Electric Interface) EEG (electroencephalogram) neural network moving average CR (conformity rate)
基金:
引用
No.****
动态公开评议
共计0人参与
勘误表
基于滑动窗口的两类运动想象EEG神经网络识别研究
评论
全部评论0/1000