互動電視之自然使用者介面設計與晶片實現

Design and Chip Implementation of Natural User Interface for Interactive Television

陳玠霖(Chieh-Lin Chen)

指導教授 ： 范育成

國立臺北科技大學/電資學院/電腦與通訊研究所/碩士(2012年)

摘要

電視的發展已經逐步地朝向智慧型電視與互動電視趨勢邁進，然而現今最通用的電視輸入裝置－遙控器，對於智慧型電視的複雜控制已經逐漸不敷使用。由於傳統的電視介面當中，只有簡單的選台、音量控制或是螢幕的色彩調整等簡單的控制，沒有額外的情況。這樣的情形在進入智慧型電視的時代後將會逐漸改變，由於智慧型電視產生了許多新的應用，因此在電視當中會有更多新的介面，若要使用遙控器來控制所有的介面，會使人感到複雜而難以使用。
本論文提出一個利用人體手部控制的自然使用者介面設計，這個系統是透過人類的手部追蹤以及手指追蹤來達到手勢辨識與手指點擊偵測等功能。透過手勢辨識與手指點擊偵測，使用者可以在操作電視時，如同在使用一個遠端的「觸控」介面，可以對電視做出特定的手勢、也能點擊電視上設計的按鈕與介面等，使得智慧型電視的介面設計更為直覺、更為人性化。此外，我們也將此系統透過硬體架構設計後，以數位積體電路設計流程實現為一顆自然使用者介面功能的數位晶片，透過讀取視訊攝影機的影像，晶片可以偵測出手部與手指的資訊與位置，做為手勢辨識與手指點擊的判斷依據。
本論文所提出的「互動電視之自然使用者介面」系統實現後，邀請了多位使用者分別參與測試，主要針對系統中的「手勢方位偵測」以及「手指點擊偵測」等兩個功能測試，所測得的平均準確率分別為97.75%及93.5%。而處理一張畫面所需的平均計算時間為18 ms，達到了即時處理的標準，使系統能夠流暢的運行。

關鍵字

自然使用者介面 ； 手指追蹤 ； 類神經網路

並列摘要

Smart TV or Interactive TV are considered as the next generation of TV set after 3D TV became the most popular options on the market. However, the interface of traditional TV, the remote control, is too simple to handle the interface of smart TVs or interactive TVs. At the same time, using the computer keyboard or mouse to control the TV was seemed a little complex or inconvenient when lying on the couch.
In this paper, we try to develop a natural user interface for interactive TV, which is more natural and easy to use with a webcam and the fingers of user. We analyze the images captured by the camera settled on the screen and track any fingers of the users. After the fingers were tracked, users can control the TV by the gestures of their hands or just use their finger to type with the virtual keyboard or buttons displayed on the TV. Finally, the main function of the system will be implemented as a chip using cell-based design to make sure the system works fluently.
After the “natural user interface of interactive TV” being implemented, we invite several user to participate the testing of the system. The testing is to find out the accuracy of the gesture direction detection and finger click detection and the tested accuracy of them relatively were 97.75% and 93.5%. Finally, we calculate the processing time, which is 18 ms, of our system when processing a frame. So the system reaches real-time processing and can be run fluently.

並列關鍵字

Natural User Interface ； Finger Tracking ； Artificial Neural Network

參考文獻

[3] 丁威群，基於DTHI技術之三維立體視訊互動系統與晶片設計，碩士論文，國立台北科技大學，台北，2011。

[5] J. Segen and S. Kumar, ”Human-computer Interaction Using Gesture Recognition and 3D Hand Tracking,” ICIP International Conference on Image Processing, vol. 3, 1998 , pp.188-192.

[7] Z. Ren, J. J. Meng and J.S. Yuan, ”Depth Camera Based Hand Gesture Recognition and its Applications in Human-Computer-Interaction,” IEEE International Conference on Information, Communications and Signal Processing (ICICS), Singapore, 2011, pp. 1-5.

[8] S.S. Rautaray and A. Agrawal, ”Design of Gesture Recognition System for Dynamic User Interface,” IEEE International Conference on Technology Enhanced Education (ICTEE), Kerala, India, 2012, pp. 1-6.

[11] M. Y. Lee and W. Woo, ”ARKB:3D Vision-based Augmented Reality Keyboard," International Conference on Artificial Reality and Telexistence (ICAT 2003), Tokyo, Japan, 2003.