應用於背投影桌之虛擬鍵盤

Virtual Keyboard for a Back Projection Desk

楊惠菁(Huei-Jing Yang)

指導教授 ： 洪一平

國立臺灣大學/電機資訊學院/資訊工程學研究所/碩士(2006年)

https://doi.org/10.6342/NTU.2006.01819

摘要

我們將一個虛擬鍵盤投影在背投影桌上，並於背投影桌之上方及下方各放置一部攝影機。兩部攝影機共同合作偵測使用者的手指觸動。首先，先建立標準鍵盤與上方攝影機之間的對應關係及
下方攝影機與上方攝影機之間的對應關係。由這兩個對應關係，在標準鍵盤上事先定義好的每個鍵的位置，即可轉換到下方攝影機及上方攝影機的座標平面上。在下方攝影機座標上的一個點可以對應到上方攝影機座標上的一條線。當下方攝影機偵測到使用者的手指觸動後，即可算出它在上方攝影機座標上所對應到的線，
初步手指偵測的技術將會沿著這條對應線搜尋手指在鍵盤上可能的位置。精確手指偵測技術將會進一步確認座落於手指可能位置週圍的那些鍵，並找出使用者真正觸動的鍵。在實驗中，我們比較四個確認方法的效能，包括主要軸分析+線性鑑別分析、二維線性鑑別分析+線性鑑別分析、支撐向量法、廣義鑑別分析。最後挑選出最好的確認方法去做進一步手指確認的動作。虛擬鍵盤的整體效能也將會在實驗中描述。

關鍵字

虛擬鍵盤 ； 投影鍵盤 ； 主要軸分析 ； 線性鑑別分析 ； 二維線性鑑別分析 ； 支撐向量法 ； 廣義鑑別分析

並列摘要

We project a virtual keyboard on the desktop of the back projection desk and put two cameras in the upper side and in the lower side of the desk. The two cameras work together to detect the user's keystroke. At first, the relation between the standard keyboard and the upper camera and the relation between the lower camera and the upper camera are built. The pre-defined position for each key on standard keyboard can be transformed to the coordinates of the lower camera and the upper camera.
A point in the coordinate of the lower camera is mapped to a line in the coordinate of the upper camera. When a keystroke is detected by the lower camera, the technique of initial finger location detection will search the corresponding line in the coordinate of the upper camera to find a potential finger location. The technique of refined finger location detection will further verify the keys situated around the potential finger location
and find out the key which the user really clicks. In the experiment, we compare the performance of four verification techniques, which are PCA+LDA, 2DLDA+LDA,SVM and GDA. Finally, the best verification technique is chosen to do verification in the refined detection of finger location. The overall performance of virtual keyboard is also shown in the experiment.

並列關鍵字

Virtual keyboard ； Projected keyboard ； Principal component analysis ； Linear discriminant analysis ； 2D discriminant analysis ； Support vector machine ； Generalized discriminant analysis

參考文獻

Abstracts, 2003, pp. 712-713.

[2] C. Tomasi, A. Rafii, and I. Torunoglu, "Full-size projection keyboard for handheld

tomatic detection of human faces in color images." in FG, 2000, pp. 54-63.

[4] H. Kruppa, M. A. Bauer, and B. Schiele, "Skin patch detection in real-world images."

[5] M.-H. Yang and N. Ahuja, "Detecting human faces in color images." in ICIP (1),