作者: Adham Atyabi , Sean P. Fitzgibbon , David M. W. Powers
DOI: 10.1007/978-3-642-35139-6_20
关键词: Brain–computer interface 、 Machine learning 、 Artificial intelligence 、 Segmentation 、 Multiplication 、 Computer science 、 Task (project management) 、 Replication (computing) 、 Electroencephalography 、 Focus (computing) 、 Motor imagery 、 Pattern recognition
摘要: EEG recording is a time consuming operation during which the subject expected to stay still for long performing tasks. It reasonable expect some fluctuation in level of focus toward performed task period. This study focused on investigating various approaches emphasizing regions interest Dividing period into three segments beginning, middle and end, expectable improve overall classification performance by changing concentration training samples had better issue investigated through use techniques such as i) replication, ii) biasing, iii) overlapping. A dataset with 4 motor imagery tasks (BCI Competition III IIIa) used. The results illustrate existing variations within potential different feasibility that regions.