結合顏色刺激及閃光頻率刺激實現四類方向辨識之研究
| dc.contributor | 葉榮木 | zh_TW |
| dc.contributor | 蔡俊明 | zh_TW |
| dc.contributor | Zong-Mu Yeh | en_US |
| dc.contributor | Chun-Ming Tsai | en_US |
| dc.contributor.author | 林詠翔 | zh_TW |
| dc.contributor.author | Yung-Shiang Lin | en_US |
| dc.date.accessioned | 2019-09-03T12:15:20Z | |
| dc.date.available | 2011-8-29 | |
| dc.date.available | 2019-09-03T12:15:20Z | |
| dc.date.issued | 2011 | |
| dc.description.abstract | 近年來，腦電波常被用來幫助肢體障礙或者脊椎受損的人們作為輔助的用具，但因腦電波裝置太過龐大以及昂貴，以至於並非普及，因此本研究為開發設計一套價格低廉的線上即時大腦人機介面系統。透過腦電波訊號處理技術應用於四方向辨識之研究，提供肢體障礙人士利用大腦意識做出控制的實用輔具。本研究利用顏色以及閃光頻率刺激的視覺回授來達成四方向的辨識，其採用的電極位置包含了大腦的頂葉(Cz)以及大腦的枕葉(Oz, O1, O2)。利用收集到的腦波資料經過前處理(數位濾波器以及獨立成分分析)將其雜訊先做去除，再利用快速傅立葉轉換觀察其顏色以及閃光頻率刺激的差異性，並採用統計方法、主成分分析法以及獨立成分分析法抽取特徵，最後投入機器學習中的支持向量機，達到辨識四方向的效果。本實驗共採用3名受測者進行測試，其離線分析最高分類率可以達到77.5%，而平均也有75%；線上即時分類率平均也可以達到68.33%，已具初步的實用價值。 | zh_TW |
| dc.description.abstract | In recent years, the technology of Electroencephalography (EEG) is used for the handicapped for tool. Because the EEG equipment is expensive and bulky, the users are not very popular. This research proposes an on-line real-time Brain Computer Interface (BCI) system. We use Electroencephalography (EEG) signal processing methods for directional controls, which may be used for the handicapped for useful tool. This research uses the optical response of the stimulus of the color and frequency of light flashes as feedback to achieve directional acquisition. The electric pole locations include the frontal lobe and occipital lobe of the human brain. The EEG data is processed to filter the unwanted noise, and then the data is processed through Fast-Fourier-Transform (FFT) to observe the difference of the signals of different color and frequency. The FFT data is then analyzed through statistics, Principal Component Analysis (PCA) and Independent Component Analysis (ICA) methods to extract the features. In the end, the data is put into a Supper Vector Machine (SVM) of a machine learning to achievedirectional acquisition. This research used 3 test subjects, which the best off-line classification rate could achieve 78.3% correctness, and the mean classification rate could achieve 75.4% correctness. Then, the on-line real time classifier accuracy could achieve 68.33% correctness, which has reached a preliminary practicability. | en_US |
| dc.description.sponsorship | 機電工程學系 | zh_TW |
| dc.identifier | GN0698730267 | |
| dc.identifier.uri | http://etds.lib.ntnu.edu.tw/cgi-bin/gs32/gsweb.cgi?o=dstdcdr&s=id=%22GN0698730267%22.&%22.id.& | |
| dc.identifier.uri | http://rportal.lib.ntnu.edu.tw:80/handle/20.500.12235/97249 | |
| dc.language | 中文 | |
| dc.subject | 大腦人機介面 | zh_TW |
| dc.subject | 腦電波 | zh_TW |
| dc.subject | 主成分分析法 | zh_TW |
| dc.subject | 獨立成分分析 | zh_TW |
| dc.subject | 支持向量機 | zh_TW |
| dc.subject | Brain Computer Interface (BCI) | en_US |
| dc.subject | Principal Component Analysis (PCA) | en_US |
| dc.subject | Independent Component Analysis (ICA) | en_US |
| dc.subject | Supper Vector Machine (SVM) | en_US |
| dc.title | 結合顏色刺激及閃光頻率刺激實現四類方向辨識之研究 | zh_TW |
| dc.title | Developing a 4-direction Brain-computer interface system based on colors and flash-light stimulation | en_US |