隨著半導體製程不斷的進步，穿戴式裝置於最近幾年得以迅速地發展。智慧型眼鏡是日常生活與工作相關應用皆日漸廣泛的穿戴式裝置之一；其目前常見的使用者輸入方式主要為語音輸入或操作觸控板。語音輸入方式是藉由智慧型眼鏡上之麥克風接收使用者的語音命令，再經由智慧型眼鏡上的語音辨識程式處理後得到輸入命令。然而，語音輸入方式由於每個人說話的音調與音量不同，容易出現誤判的情況，在吵雜的環境中更是不易於使用。觸控板操作輸入方式是藉由智慧型眼鏡側邊之觸控板來接收使用者的輸入。此種輸入方式使用上較不直覺且觸控面積小，與擁有觸控螢幕的智慧型手機其操作體驗相比之下，使用起來較不流暢且觸控靈敏度顯得較低，若在使用者配戴手套的情況下就無法使用。綜合上述原因，本論文研究設計並實作一套適用於智慧型眼鏡之懸浮手勢操作介面。
本論文實現之懸浮手勢操控功能，提供給智慧型眼鏡之使用者另一種操控方式。此智慧型眼鏡懸浮手勢操控功能以Android應用程式的形式實現，可運作在基於Android系統之智慧型眼鏡，並具有下列特點：(1)適用於智慧型眼鏡之使用者介面、不受限於單一或少數應用程式；(2)使用單一相機擷取畫面並經由影像處理完成手勢辨識；(3)使用者無須穿戴任何感應裝置，便可直接透過手勢操控智慧型眼鏡；(4)模擬使用者對於觸控板之輸入，易於與現有應用程式整合；(5)手勢辨識流程模組化設計，方便開發者對辨識流程進行客製化。此套人機互動介面平均單次操控可於0.5秒內完成，除了應用於一般操作，也可擴充作為未來智慧型眼鏡對擴增實境與混合實境之操作方式的概念雛形。

As wearable technology has been advancing fast over the last few years, smart glasses are getting prevalent nowadays and they are used in many different ways in our daily life and work. The common input methods for smart glasses are voice input control and touchpad control. The voice input control is liable to errors due to different tones and volume variation. Furthermore, inaccurate recognition due to the background noise in the environment might occur. The touchpad control is not user-friendly due to its size restriction on the smart glasses. The area of the touchpad is often too small to provide a pleasant user experience. Both voice input control and touchpad input control have cons.

In order to provide an alternative input method for the users of smart glasses, this thesis attempts to realize a feature which is called air gesture control. With this feature, user can execute control functions without touching the touchpad. Users can do the common basic operations, such as swiping left or right, scrolling and going to the home page, just by making hand gestures in front of the built-in camera of the smart glasses. The air gesture control is realized by implementing a hand gesture recognition program which consists of the following: (1) Pre-processing Module, (2) Image Enhancement Module, (3) Handshape Extraction Module, (4) Classification Module, and (5) Output Module. The average time of recognizing a gesture is less than 0.5 second.
The main contribution of this thesis is providing an alternative input method for the users of smart glasses. The feature has the following characteristics: (1) It is designed for general input method of Android system rather than a particular application; (2) It accomplishes hand gesture recognition by single camera; (3) Users can control the smart glasses via hand gesture without wearing any sensing aid; (4) It simulates the touchpad input events and can be easily integrated with other applications; (5) The modularization of the hand gesture recognition program is helpful for the developers to customize their programs. This style of human-machine interfaces can be regarded as basis of smart glasses’ control method for applications of augmented reality and mixed reality.

[1] Research and Markets, “Smart Glass Market, 2023,” https://www.prnewswire.com/news-releases/smart-glass-market-2023-300534726.html, accessed on 2019-02-18.
[2] Google, Google Glass, https://www.google.com/glass/start/, accessed on 2019-06-19.
[3] MacRumors staff, Apple Glasses, https://www.macrumors.com/roundup/apple-glasses/, accessed on 2019-09-03.
[4] Microsoft, Microsoft HoloLens, https://www.google.com/glass/tech-specs/, accessed on 2018-06-19.
[5] Vuzix, Vuzix M300, https://www.vuzix.com/support/M300-smart-glasses, accessed on 2018-06-19.
[6] Epson, Moverio BT-300, https://tech.moverio.epson.com/en/bt-300, accessed on 2018-06-19.
[7] Sony, SmartEyeglass, https://developer.sony.com/develop/smarteyeglass-sed-e1/, accessed on 2018-06-19.
[8] Jorjin, Jorjin J-100, https://www.jorjin.com/product.php?id=177, accessed on 2018-06-19.
[9] Roger Schaer, Thomaz Melly, Henning Muller, Antoine Widmer, “Using smart glasses in medical emergency situations, a qualitative pilot study, ” IEEE Wireless Health (WH), 2016.
[10] 資訊工業策進會, “智慧眼鏡工業應用解決方案,” https://www.iii.org.tw/Product/TechLensDtl.aspx?tp_sqno=h2bypOe0AoHFu7mVYuTK5A__&fm_sqno=55, accessed on 2019-06-30.
[11] Mark Prigg, “The app that lets you shoot round corners: Smart rifle uses Google Glass as a remote sight,” http://www.dailymail.co.uk/sciencetech/article-2649051/The-app-lets-shoot-round-corners-Smart-rifle-uses-Google-glass-sight.html, accessed on 2018-06-19.
[12] Microsoft, Kinect, https://en.wikipedia.org/wiki/Kinect, accessed on 2018-06-19.
[13] BMW, BMW gesture control, https://www.bmwgroup.com/en/innovation/technologies-and-mobility/models.html, accessed on 2018-06-19.
[14] Natural user interface, https://en.wikipedia.org/wiki/Natural_user_interface, accessed on 2018-06-19.
[15] Rick Merritt, “Smart Glasses Not Ready to Wear,” https://www.eetimes.com/document.asp?doc_id=1332879, accessed on 2018-06-19.
[16] Milind R. Hegade, Neha Jamdar, Abhijit Nale, Ankush Barkade, Pravin Ekhande, VIShreyas Dayal,” Hand Gesture Recognition : A Survey,” International Journal of Advanced Research in Computer Science & Technology, 2015.
[17] Data Glove, http://fifthdimensiontechnologies.com/5dt-data-glove-ultra, accessed on 2019-02-18.
[18] J. Brand, J. S. Mason, “A comparative assessment of three approaches to pixel-level human skin-detection,” Proceedings 15th International Conference on Pattern Recognition. ICPR-2000, 2000.
[19] M. J. Jones, J. M. Rehg, “Statistical color models with application to skin detection,” Proceedings. 1999 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No PR00149), 1999.
[20] B. D. Zarit, B. J. Super, F. K. H. Quek,” Comparison of five color models in skin pixel classification,” Proceedings International Workshop on Recognition, Analysis, and Tracking of Faces and Gestures in Real-Time Systems. In Conjunction with ICCV'99 (Cat. No.PR00378), 1999.
[21] Son Lam Phung, A. Bouzerdoum, D. Chai, “ A novel skin color model in YCbCr color space and its application to human face detection,” Proceedings. International Conference on Image Processing, 2002.
[22] OpenCV, OpenCV org, https://opencv.org/, accessed on 2018-06-19.
[23] OpenCV, OpenCV wiki, https://github.com/opencv/opencv/wiki, accessed on 2018-06-19.
[24] OpenCV, https://en.wikipedia.org/wiki/OpenCV, accessed on 2018-06-19.
[25] 王曦, “智慧型眼鏡一戴 20 年：專訪 Google Glass 技術負責人 Thad Starner,” https://www.inside.com.tw/2014/02/07/google-glass-thad-starner-interview, accessed on 2018-06-19.
[26] Bradley Rhodes,” A brief history of wearable computing,” https://www.media.mit.edu/wearables/lizzy/timeline.html#1981b, accessed on 2018-06-19.
[27] Ben Popper, “New evidence emerges in alleged assault on cyborg at Paris McDonald's,” https://www.theverge.com/2012/7/19/3169889/steve-mann-cyborg-assault-mcdonalds-eyetap-paris, accessed on 2018-06-19.
[28] 張岑宇, “「穿戴式裝置之父」Steve Mann　頭戴智慧眼鏡超過35年被當怪咖,” https://www.limitlessiq.com/news/post/view/id/1267/, accessed on 2018-06-19.
[29] Android, https://en.wikipedia.org/wiki/Android_(operating_system), accessed on 2018-06-19.
[30] Scott Stein, “Google Glass returns: This time, it’s professional,” https://www.cnet.com/news/google-glass-2-goes-for-enterprise, accessed on 2019-02-18.
[31] Android Platform Architecture, https://developer.android.com/guide/platform#api-framework, accessed on 2019-04-26.
[32] Java native interface, https://docs.oracle.com/javase/8/docs/technotes/guides/jni/, accessed on 2019-06-30.
[33] Intermediate representation, https://en.wikipedia.org/wiki/Intermediate_representation, accessed on 2018-06-19.
[34] Dalvik, https://en.wikipedia.org/wiki/Dalvik_(software), accessed on 2018-06-19.
[35] Jonathan Levin, Android Internals:Power User's View, http://newandroidbook.com/, accessed on 2018-06-19.
[36] Samsung, “What is the Air Gesture feature and how do I use the Air Gesture feature?,” https://www.samsung.com/hk_en/support/mobile-devices/what-is-the-air-gesture-feature-and-how-do-i-use-the-air-gesture-feature, accessed on 2019-02-18.
[37] GT-19505 User Manual, https://www.samsung.com/uk/support/model/GT-I9505DKYORO, accessed on 2019-02-18.
[38] Wave Control, https://play.google.com/store/apps/details?id=com.MarksThinkTank.WaveControl, accessed on 2019-02-18.
[39] Air Gesture Control – PRO, https://play.google.com/store/apps/details?id=in.tank.corp.proximitypro, accessed on 2019-02-18.
[40] Air Call Receiver, https://play.google.com/store/apps/details?id=com.aircall.reciever&hl=en_US, accessed on 2019-02-18.
[41] Francesca Gasparini, Raimondo Schettini, “Skin segmentation using multiple thresholding,” Proceedings Volume 6061, Internet Imaging VII, 2006.
[42] Kyung-Sun Lee, Myung-Chul Jung, “Flexion and Extension Angles of Resting Fingers and Wrist,” International Journal of Occupational Safety an Ergonomics, vol.20, no.1,pp.91-101, 2014.
[43] Sendevent, https://android.googlesource.com/platform/system/core/+/android-5.0.2_r3/toolbox/sendevent.c, accessed on 2019-02-18.
[44] Touch devices, https://source.android.com/devices/input/touch-devices, accessed on 2019-02-18.
[45] Multi-touch protocol, https://www.kernel.org/doc/Documentation/input/multi-touch-protocol.txt, accessed on 2019-02-18.
[46] TimingLogger, https://developer.android.com/reference/android/util/TimingLogger, accessed on 2019-02-18.
[47] Kai-ping Feng, Fang Yuan, “Static hand gesture recognition based on HOG characters and support vector machines,” 2013 2nd International Symposium on Instrumentation and Measurement, Sensor Network and Automation (IMSNA).