基于典型相关分析的眼动辅助脑电假手动作识别方法研究  被引量:1

Recognition Method for Eye-moving Assisted Neuroprosthetic Hand Movement Based on Canonical Correlation Analysis

在线阅读下载全文

作  者:孙曜[1,2] 罗志增[2] 韦巍[1] 

机构地区:[1]浙江大学电气工程学院,浙江杭州310027 [2]杭州电子科技大学机器人研究所,浙江杭州310018

出  处:《航天医学与医学工程》2011年第2期128-133,共6页Space Medicine & Medical Engineering

基  金:国家自然科学基金项目(60874102);浙江省公益项目(2010C33131);浙江省教育厅科研项目(Y200909010)

摘  要:目的解决采用单纯运动想象方式难于获得具有明显特征的脑电信号(EEG)进而导致假手动作识别困难的问题。方法利用伴随大脑思维过程的自然眼部运动所引发的眼电信号增强运动想象脑电信号特征,获得便于识别的眼动辅助运动想象脑电信号。通过对类内样本集及类间样本集进行典型相关分析来选取模式识别的特征变量;再提取待识别信号集与各模式样本信号集间作为特征变量的典型变量,依据其相关性强弱进行分类识别。结果应用所提出方法进行假手6种动作识别所获得的平均识别率为87.6%,远高于基于单纯运动想象脑电信号进行对比实验所获得的识别率。结论眼动辅助运动想象脑电信号较之单纯运动想象脑电信号,具有更大的类间差异,更强的类内相似度,应用本文方法能显著提高脑电假手的动作识别率。Objective To overcome the difficulty in recognizing neuroprosthetic hand movement electro-encephalogram(EEG) patterns due to the hardness to obtain EEG with obvious features only by pure motor imagination(MI).Methods The easily recognizable eye-moving assisted motor imagery(EAMI) EEG was obtained by utilizing the electro-oculogram(EOG) caused by natural eye movement associated with the brain thinking process to enhance the MI EEG features.The feature variable used for pattern recognition was not rashly chosen.The suitable canonical correlative variable was chosen as the feature variable with canonical correlative analysis(CCA) to analyze the correlation of the intra-class samples and inter-class samples sequentially.The feature variables between the need-to-recognize signal set and each class sample signal set were extracted to make classified recognition according to degree of correlation.Results Average recognition rate of 87.6% was obtained with this proposed method to recognize six motion patterns of neuroprosthetic hand based on EAMI EEG.It was much higher than the average recognition rate obtained in the contrast experiment based on pure MI EEG.Conclusion Compared with pure MI EEG,the EAMI EEG is of more inter-class difference and more intra-class similarity.This proposed method can improve the motion pattern recognition rate of neuroprosthetic hand significantly.

关 键 词:脑电图 眼电图 假手 典型相关分析 

分 类 号:R857.2[医药卫生—航空、航天与航海医学]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象