摘要
脑电癫痫特征波的自动提取对于患者的诊断具有重要的意义。提出一种时频分析与Jensen函数相结合的方法进行棘波检测,然后提取出棘波的波形特征,并通过人工神经网络进行进一步的判决,从而降低棘波检测的误检率。在对真实的癫痫脑电信号(EEG)的仿真实验中,该方法取得了较好的结果。
The automatic spike detection in EEG is significant in diagnosing neural diseases. This paper proposed a new method which combined time-frequency analysis and Jensen function to detect spike automatically. The characteristics of the detected spikes were extracted with an artificial neural network（ANN） to make further judgement to reduce the false detection ratio. Applying the method to real epileptic EEG, expected results were obtained.
出处
《中国生物医学工程学报》
CAS
CSCD
北大核心
2006年第4期421-425,429,共6页
Chinese Journal of Biomedical Engineering
基金
国家自然科学基金资助项目(30170259
30570475)
辽宁省科学技术基金资助项目(2001101057)
关键词
脑电信号
时频分析
人工神经网络
棘波检测
EEG
time-frequency analysis
artificial neural network
spike detection
