身心障礙者於日常生活中有諸多不便，尤其全身癱瘓者，幾乎所有日常生活瑣事皆需他人幫忙，且生活極為枯燥乏味。有鑒於此，本實驗室發展出一輔助性電腦輸入裝置，提供身心障礙者操控嘴控開關，以摩斯碼輸入方式來控制鍵盤及滑鼠。另外也發展出一套輔助性家電控制系統，使用者可透過電腦操控家電，包含電視、冷氣、…等等。然而使用者仍需由它人幫忙將嘴控開關放置於患者嘴上。因此本研究提出一套自動奶嘴開關載運系統，讓患者不須經由它人幫忙，就可自行取用或收回奶嘴開關，達到系統自動化，更提高身心障礙者使用整個系統的完整性。
系統硬體架構採用樂高積木組合而成，可減少硬體原型開發的時間，軟體則是採用LabVIEW影像處理軟體。本系統採用影像定位方式，以市售網路攝影機來擷取使用者臉部之影像，先將影像轉換到色相空間取得人臉膚色的區域，接著判別膚色區域是否符合人臉比例，再以眼睛約在人臉的1/2處為條件，藉此以縮小樣本比對之範圍，以達到人眼位置之定位，最後利用眼睛與嘴巴的位置相對關係，將奶嘴開關準確送至使用者的嘴巴內。
實驗室初步測試系統平均成功率為90%，平均辨識相似度為97.6%，平均每次耗費時間為36.83秒，但於個案測試時卻因測試場地不同而使得成功率降低很多，調整桌面水平後辨識成功率約為20%。此系統之可行性及實用性仍有待改善，另外辨識速度也為後續需研究可改進之處。

There are many inconveniences associated with having disabilities, especially for quadriplegics. The necessity for help in basic daily activities contributes to their monotonous and dismal outlook on life. For these reasons, our research group has developed an assistive computer input device to help the disabled control a keyboard and mouse by a nipple switch with Morse code encoding. In addition, an assistive home appliance control system controlled through the computer was developed to help the handicapped control appliances by themselves including TV, AC, and other electrical devices. However, before operating the computer, the assistive computer input device has to be placed into the mouth of the user by another person. Therefore, we propose an automatic transportation system to position/remove the patient's nipple switch, making the system considerably more automated and convenient.
The hardware of the system was constructed with Lego, while the software was developed with the image process software of LabVIEW, which greatly enhanced the developmental process. The positioning method of the system is via image processing.
The human facial image was captured by a marketable webcam. First, the facial area was obtained by using the transformation of colorful space. Next, we distinguished the proportionality of the facial area, making sure that the eyes are located within the top 2/3 of the face to reduce the area of the pattern recognition. Afterwards, we obtained the position where the eyes were located. Finally, the nipple switch was placed into the user's mouth according to the relative location of the eyes and mouth.
In our experiments conducted in the laboratory, the average accuracy of the proposed system was 90%, the recognition rate was 97.6%, and the average time of transporting the nipple switch to the mouth was 36.83s. In the practical experiments, the average accuracy of the proposed system decreased considerably to 20% due to environmental variability. The practicability of the proposed system was not sufficient and the recognition speed was too slow. Both points will have to be improved in future research.

[1] 吳崇民著，"應用於重度及脊髓損傷患者之摩斯碼模糊控制辨識嘴控系統" ，國立成功大學電機工程學系博士論文，中華民國93年11月。
[2] 張慶龍著，"超音波馬達定位輔助系統" ，國立成功大學電機工程學系碩士論文，中華民國95年6月。
[3] 邵成麟著，"以LabVIEW與IMAQ Vision為基礎實現之唇形影像辨識溝通輔具系統"，南台科技大學電機工程研究所。
[4] 張興政著，"The Dynamic Sensing and Analyzing Platform Base on 3D movements of a Robotic Arm"，逢甲大學資訊電機工程學系碩士論文，中華民國97年1月。
[5] 李宜承撰，"有效向量法則在高自由度機器人上的改良與應用研究"，逢甲大學自動控制工程學系碩士論文，中華民國97年5月。
[6] 劉素妹撰，"以樂高積木為媒介提升幼兒口語能力之研究"，國立台東大學幼兒教育學系碩士論文，中華民國97年8月。
[7] 王健宇撰，"可變形樣板於即時人臉追蹤系統之應用"，國立中山大學機械與機電工程學系碩士論文，中華民國96年7月。
[8] 陳宣銘撰，,"Non-Photo-Realistic Illustrations with Artistic Style"，國立中央資訊工程學系碩士論文, 中華民國92年12月。
[9] Baogang Wei, Yonghuai Liu, Member, IEEE, and Yunhe Pan，"Using Hybrid Knowledge Engineering and Image Processing in Color Virtual Restoration of Ancient Murals"，IEEE Transactions On Knowledge and Data Engineering, vol. 15, no. 5, September/October 2003.
[10] 曾郁展著，"利用人臉五官為特徵之人臉辨識系統"，國立中山大學電機工程學系碩士論文，中華民國94年6月。
[11] 紀智超著，"中醫舌診電腦化舌下絡脈特徵擷取模糊分析"，國立中山大學應用數系碩士論文，中華民國88年6月。
[12] NI-IMAQ for USB Cameras User Guide"2005 edition National Instruments Corporation.
[13] Thomas Heseltine1, Nick Pears and Jim Austin，"Evaluation of image pre-processing techniques for eigenface based face recognition"，The Proceedings of the Second International Conference on Image and Graphics, SPIE vol. 4875, pp. 677-685 (2002).
[14] Peter N. Belhumeur, Joa~o P. Hespanha, and David J. Kriegman, "Eigenfaces vs. Fisherfaces: Recognition Using Class Specific Linear Projection"，IEEE Transactions on Pattern analysis and Machine intelligence, vol. 19, no. 7, July 1997.
[15] 羅竣威、謝清祿著," 三種主軸成分法於數位影像特徵擷取與判別之研究"， 國立屏東科技大學機械系碩士論文。
[16] Wangmeng Zuo, David Zhang, Senior Member, IEEE, and Kuanquan Wang, Member, IEEE，" Bidirectional PCA With Assembled Matrix distance Metric for Image Recognition "，IEEE Transactions on Systems, Man, and Cybernetics-Part B: Cybernetics, vol. 36, no. 4, August 2006.
[17] Turk, M., Pentland, A., 1991, Eigenfaces for Recognition. Journal of Cognitive Neuro-science, 3(1), pp.71-86.
[18] LabVIEW QuickStart Guide National Instruments 1998 January Edition
[19] LabVIEW user manual 2001 edition National Instruments Corporation。
[20] LabVIEW Toolkit for LEGO Mind storms NXT Programming Guide 2006 edition National Instruments Corporation.
[21] 林朝枝、葉建廷、劉木坤著，"An algorithm for solving convex hull problem 向外凸殼問題的演算法" ，吳鳳科技大學資訊工程系。
[22] 陳志強著，"影像處理技術應用於PLED噴墨基板瑕疵檢測"，中華民國92年7月。