Wavelet and Common Spatial Pattern for EEG Signal Feature Extraction and Classification
Brain-computer interface (BCI) can provide communication channels which do not depend on peripheral nerves and muscles for patients with neuromuscular disorders. The goal of the paper is to validate signal processing and classification methods for Brain-Computer Interfaces (BCIs). This paper presented a method combining wavelet with Common Spatial Pattern (CSP). Use multi-resolution analysis (MRA) to weaken noise and enhance features in the signals of motor imagery electroencephalogram (EEG). Then features are extracted and classification is completed by Common Spatial Pattern and Support Vector Machine (SVM) respectively. The classification accuracy achieves 93.5％ in the course of testing on the data from subject. The result certifies the feasibility and effectiveness of this solution.
brain-computer interface wavelet transform common spatial pattern support vector machine
Liwei Zhang Guozhong Liu Ying Wu
School of Photoelectronic Information & Communication Engineering Beijing Information Science & Technology University Beijing, China
国际会议
长春
英文
242-246
2010-08-24（万方平台首次上网日期，不代表论文的发表时间）