Incorporation of Neural Networks and HHT for Recognition of EEG Signals

瀏覽統計 Email 通知 RSS Feed

簡歷

基本資料

Project title

Code/計畫編號

NSC98-2221-E019-032

Translated Name/計畫中文名

結合神經網路與HHT於腦波信號之判讀

Project Coordinator/計畫主持人

Jung-Hua Wang

Funding Organization/主管機關

National Science and Technology Council

Department/Unit

Department of Electrical Engineering

Website

https://www.grb.gov.tw/search/planDetail?id=1914997

Year

2009

Start date/計畫起

01-08-2009

Expected Completion/計畫迄

31-07-2010

Bugetid/研究經費

630千元

ResearchField/研究領域

資訊科學--軟體
醫學工程

腦波訊號之頻率關聯性或時間-空間特性係非穩態(non-stationary)、非線性，具高變異性，不易精確偵測EEG信號瞬時事件(transient event)，本計畫擬使用希爾伯特-黃(HHT) 轉換技術擷取腦波訊號之時頻特徵向量作為訓練神經網路（Neural Network）之輸入向量，配合神經網路的學習能力，將以三個不同spatio-temporal之神經網路模型，搭配前述HHT 時頻特徵資料，並以預先建立的驗證資料庫來進行訓練測試，各別網路訓練完畢後，再施以一微調訓練，即允許不同神經網路間互相學習，相較於傳統訓練單一神經網路，可更有效增加神經網路之概泛能力進而降低EEG高變異性的影響以獲得最適網路架構(層數及各層神經元數目)。預期由於神經網路對於雜訊有一定的強建性，因此依據輸入多通道腦波信號，學習後的神經網路可作為一個強建性分類器(classifier)或腦波異常量化器。採用神經網路於偵測並判別異常波訊號，其特色係不須假設特徵向量機率密度函數之型式為已知，藉由訓練神經網路學習偵測辨識腦波訊號，輔助癲癇病症診斷，提升醫療品質。This research project aims to develop a technique useful in detecting multi-channel EEG transient events or characteristic waves, such as spikes, spindle, hump, etc. Our approach is based on utilizing a set of Hilbert-Huang Transformation (HHT) feature data to train three different spatio-temporal neural networks. Key essence of the training process is it allows inter-learning among the three neural networks in order to further enhance the generalization capability of the three neural networks for building a characteristic waves classifier or transient event detector.

Keyword(s)

類神經網路
希爾伯特-黃轉換
圖樣識別
分類器
Neural Network
Hilbert-Huang Transformation
Pattern Recognition
Classifier

DSpace CRIS

Incorporation of Neural Networks and HHT for Recognition of EEG Signals

基本資料

Description