溝通對於一個意識清醒卻無法正常表達的漸凍人而言，是一件遙不可及的事。倘若有一套不同以往的溝通介面，在使用者對外界表達的一般功能完全喪失的情形下，能夠很清楚傳達他的意思。那對於漸凍人的對外界溝通與病情了解方面，將有莫大的幫助。 本研究的目的在於建構一套“眼動符號表達系統”，就是使用眼睛特定移動的軌跡來表達特定符號的系統。藉由量測眼睛周圍微弱的電位差，包含雙眼左右與單眼上下的電位差，以得知眼睛移動的方向，並使用電腦軟體處理分析技術來判定使用者所表達的符號。 在本論文中，以眼電圖法 (Electro-OculoGraphy, EOG) 來作為擷取眼睛移動訊號的方法，並使用眼動訊號震盪的方向、次數與相關性作為學習向量量化 (Learning Vector Quantization, LVQ) 類神經網路的輸入特徵值，藉此判斷使用者所勾勒出的符號。系統裡共內建了10個數字和4個四則運算符號，讓使用者傳達不同的意思。 Communicating is an impossible mission for disables who have normal function eyes and is conscious, but can’t express normally. If there exists a new communicating interface, in a situation of users express function loss completely, it could transmit his will clear. It is utmost help for the outside communicating to disables and patients' condition understanding. Our research objective is constructing a unit of “Symbols Expression System Based On Eye Movement”, that is a system that utilize particular eye movement locus to express particular symbols. By sensing week potential around eye to be aware of the direction of eye movement and using software analysis technology to judge the users express symbols. In my thesis, I utilize Electro-OculoGraphy (EOG) to be the method of acquiring eye movement signal. Using the eye shaking direction, the number of times and correlation as the input characteristic value of the Learning Vector Quantization (LVQ) Neural Network, depend it judging the symbols draw by user. In the system, there are fourteen kinds of symbols for use, hence, it could allow user to express different mean.
