跳到主要內容區塊

 

103 腦電波情緒識別系統之開發研究

出版年度:103

 

論文名稱:腦電波情緒識別系統之開發研究

 

研究生:施韋良   指導教授:蔡子瑋

 

 


論文簡介:

  透過腦電波生理訊號理解人們的情緒反應, 近年來感性人因工程 及神經行銷學, 是逐漸重視的研究議題, 然而腦電波對情緒分類的影 響, 及以腦電波為基礎的情緒識別系統的研究尚為欠缺。因此, 本研 究目的在於: 1.評估腦電波訊號對情緒分類的影響; 2.開發一以腦電 波為基礎的情緒識別系統。本研究使用NeuroSky 單通道腦波儀, 並 將腦電波訊號資料以J48 決策樹演算法進行分類, 研究結果顯示對情 緒分類有影響的腦電波訊號:At tent ion、Medi tat ion、Theta、Low alpha、High alpha、Low beta、High beta、Low gamma、Mid gamma, 而腦電波訊號Del ta 與情緒影響度較低。此外, 透過決策樹分類情緒的組織結構建置的腦電波情緒識別系統, 可分類出四類情緒: 悲傷、愉悅、厭惡、額外情緒, 此系統經實驗驗證準確率為63%。 In recent years, the research of Kansei Engineering and Neurological market ing are get t ing important and valued a lot . However, i t is st i l l an open quest ion how the brainwave influent emot ion ident ificat ion and how to ident ify emot ion from brainwave response. Thus, the aims of the thesis are: 1.) to explore the effect of people’s brainwave on emot ion classificat ion; 2.) to develop the emot ion ident ificat ion system based on brainwave. The single-channel electroencephalograph developed by NeuroSky Co. is adopted to record part icipants ’ brainwaves. The brainwave singles are computed wi th J48 decision tree algori thm to resul t the emot ion classificat ion. Then, the emot ion classificat ion resul ts are appl ied to develop an emot ion ident ificat ion system which is evaluated afterward. The conclusions yield that the brainwaves wi th significant effects on emot ion classificat ion are At tent ion, Medi tat ion, Theta, Low alpha, High alpha, low beta, High beta, Low gamma and Mid gamma, whi le the Del ta has less significant effects on emot ion classificat ion. Moreover, the emot ional ident ificat ion system based on brainwave could be appl ied to classify four one’s emot ions: sadness, pleasure, disgust , reject ion emot ion. The evaluated accuracy reveals 63%.

瀏覽數