结合CSP 和HMM 的左右手运动想象分类
宋俊可，吴小培
基金项目：博士点基金（200803570002）；国家自然科学基金资助项目(60771033)
作者简介：宋俊可，（1988-），女，硕士研究生，生物医学信号处理.
（安徽大学，计算智能与信号处理教育部重点实验室，合肥 230039）
摘要：脑机接口是不依赖于动作和语言，在大脑与计算机或其它电子设备之间建立直接的交流和控制通道，让人可以通过大脑来直接的表达想法和操纵设备。对于以运动想象为基础的脑机接口系统，文中把2003BCI 标准数据和实验室实测数据作为处理对象，将数据经过分段带通滤波预处理、共同空间模式（CSP）空间滤波提取特征向量，最后用隐马尔科夫模型（HMM）进行分类识别。实验结果表明，该方法是一种有效的运动想象分类方法。
关键词：脑机接口；运动想象；共同空间模式；隐马尔科夫模型
中图分类号：R318
Classification of Right and Left Hand Motor Imagery Based on CSP Algorithm and HMM
SONG Junke, WU Xiaopei
(The Key Laboratory of Intelligent Computing & Signal Processing Anhui University,HeFei 230039)
Abstract: Brain-Computer Interface（BCI）enables the users to express their ideas and manipulate devices through their brains directly by creating an interaction and control channel between human brain and computers or other electronic equipments without any language speaking or body movement. In this paper, based on a motion imagination BCI system, CSP
spatial filtering is adopted to extract features after a band-pass filtering as a preprocessing is performed. The classification is finally achieved using HMM. The experiment results on both the 2003 BCI standard data and the real-life data collected by our laboratory demonstrate that the proposed algorithm works effectively in motor imagery classification.
Keywords:Brain-Computer Interface(BCI); motor imaery; Common Spatial Pattern(CSP); Hidden Markov Model(HMM)