Please use this identifier to cite or link to this item:
標題: 在觸控螢幕手勢辨識之設計與實作
Design and Implementation of Gesture Recognition on TFT
作者: Yen-Wen Chiu
關鍵字: gesture recognition
引用: 參考文獻 [1] DisplaySearch, '平面顯示面板市場-2012年度回顧及2013年預測', [Online]. [2] Wikipedia, 'TFP-LCD',[Online]. [3] 平顯資訊, 'TFT-LCD結構',[Online]. [4] Pentileblog, 'RGB STRIPE AS A BENCHMARK',[Online]. [5] 洪光煜, 'Horizontal Synchronize Signal', 義守大學電子工程研究碩士學位論文, 2006 [6] NextWindow, 'Optical Touch: How it differs from other technologies',[Online]. [7] G Touch, '觸控原理',[Online]. [8] Jonathan design, '觸控面板感應原理',[Online]. [9] 楊瑞賢, '觸控面板技術介紹',2009,[Online]. [10] 粘存賢, '設計與實作一個基於雙攝影機視覺系統的雙點觸控螢幕', 中央大學資訊工程學系碩士學位論文, 2009. [11] Display All, '各種觸控方式的原理與特徵', 2009,[Online]. [12] Jonathan design, '投射式電容觸控原理', 2010,[Online]. [13] 新通訊, '感測電路結構各展所長 電容式觸控技術百家爭鳴', 2008, [Online]. [14] Hao Sun and Kaiyu Shen, 'Virtuoso: A Touchscreen Music App', 2012, [Online]. [15] 'The Center for New Music and Audio Technology', [Online]. [16] 'Open Sound Control', [Online]. [17] 'TUIO', [Online]. [18] Patently Apple, 'Apple Wins Strategic Multitouch & Music Tempo Workout Patents', 2010, [Online]. [19] 陳慶翰, 'GRAFCET離散事件建模', 中央大學資工系教程, 2009. [20] 陳慶瀚, 'PNN機率式神經網路', MIAT實驗室, 2011. [21] Zaknich, 'Introduction to the Modified Probabilistic Neural Network for General Signal Processing Applications', IEEE Trans. On Signal Processing, Volume 46. No. 7, pp. 1980-1990. , 1998. [22] D. F. Specht, 'Probabilistic Neural Networks (original contribution)', Neural Networks, Volume 3, No.1, pp.109-118. , 1990. [23] D. F. Specht, 'Probabilistic Neural Networks (original contribution)', Neural Networks, Volume 3, No.1, pp.109-118. , 1990. [24] T. N. E. Greville, 'Theory and Application of Spline Functions', New York: Academic, 1969 [25] 浩宇電子, '3.5寸TFT[320*240]24位色LCD顯示模組', 2013, [Online]. [26] 浯陽科技, 'MIAT-STM32 嵌入式ARM Cortex M3開發平台', 2013, [Online].
摘要: Smart touch screen which is the most important connection with users made human-computer interaction the most popular universal usage. And gesture recognition e.g., left, right, up and down, had become necessary. In this thesis, we design and implement a real-time gesture recognition system. On the basis of the probabilistic neural network, that is a kind of classifier gesture modeling, we first build some gesture models. Then, uponing receiving a series of touching points, we filters abnormal points, and then perform gesture point coordinates normalization and linear interpolation. Finally, we compared the models established by PNN classifier for the classification of gesture recognition. We realized about gesture recognition on the resistive touch panels. Besides, we can increase the training data on the PNN classifier model with different users and different gesture to improve the gesture recognition rate. Experimental results show that our system can immediately and successfully recognize gestures. Besides, due to our simplified system design, our system can be used in different embedded systems with low power pan-touch device.
現今人機互動普及,智慧型觸控螢幕是最重要的一環,其中系統操作上的手勢辨識如向上、下滑動,向左前頁、向右後頁的手勢辨識,已成了操作上的直覺配備。本研究提出一套整合觸控顯示的手勢辨識系統,透過以機率神經網路分類器(Probabilistic Neural Network PNN)建立簡易模組後,在以不增加系統負擔的前提下,由系統即時將操作座標經過異常點座標濾除、座標正規化及座標插補後與分類器所建立的模組比對分類,實現以薄膜電晶體液晶顯示器TFT-LCD為操作介面,進行系統軟體的手勢辨識。此系統可在建立機率神經網路分類器時,增加由不同來源的使用者手勢的訓練母數,提高正確的辨識率。由最後實驗結果,此一方式可即時辨識操作手勢,並因簡化辨識系統的設計下,可用於不同的觸控操作裝置以實現低功率嵌入式系統泛用性功能。
其他識別: U0005-2008201515235600
文章公開時間: 10000-01-01
Appears in Collections:資訊科學與工程學系所



Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.