應用於簡報作業之人機智慧型互動技術之研究

An Intelligent Interactive Interface For The Presentation System

王富民(Fu-Min Wang)

指導教授 ： 賴薇如

元智大學/電機通訊學院/通訊工程學系/碩士(2012年)

摘要

本論文提出一套經由分析微軟Kinect設備輸入的深度影像，擷取出演講者的手勢，以控制簡報作業之人機智慧型互動系統。我們透過深度影像擷取人體骨架化（Skeleton）資訊，同時使用深度影像和骨架化資訊擷取出完整的手部區域進行動作與手勢分析。動作識別是透過手部3D軌跡做為辨識依據；手勢識別則是透過手指的輪廓與掌心的距離和角度轉換成手指曲線做為特徵，再利用Finger-Earth Mover’s Distance （FEMD）演算法來判斷手勢。我們將動作和手勢應用於簡報的控制上，盡可能擺脫鍵盤與滑鼠的限制，達成人機互動的目的。

關鍵字

微軟體感攝影機 ； 人機互動 ； 人體骨架化 ； 手勢

並列摘要

The thesis brings up an idea of using the depth image and human skeleton information from Microsoft Kinect to control the Microsoft PowerPoint. That is, we create an intelligent human-computer interaction system. The system acquires the motion of a performer via Kinect-based human skeleton tracking, and analyses 3D trajectory to control the PowerPoint. On the other hand, we try to get complete hands areas by using depth image from Kinect. However, we get the finger curve which records the relative distance between each contour vertex to a center point of hand, and then, use the Finger-Earth Mover’s Distance algorithm to recognize hand gestures. We apply the result of action recognition and by gestures to control the PowerPoint, and thus we can break away from the restriction of keyboards and mice to achieve the goal of the human-computer interaction.

並列關鍵字

Kinect ； Human-Computer Interaction ； Skeleton ； Hand Gesture ； FEMD

參考文獻

[30] 林玉瓊, et al., "十分鐘走路熱身運動對全反應時間不同分段時間值之影響," 體育學報, pp. 293-304, 1995.

[31] 林淑親 and 林耀豐, "影響反應時間因素之探討," 中華體育季刊, vol. 21, pp. 103-117, 2007.

The Effects of warm-up and static stretch on reaction time and movement time of the elementary children," 碩士, 運動健康研究所.

[3] D. J. Sturman and D. Zeltzer, "A survey of glove-based input," IEEE Computer Graphics and Applications, vol. 14, pp. 30-39, 1994.

[5] Z. Ren, et al., "Robust hand gesture recognition based on finger-earth mover's distance with a commodity depth camera," presented at the Proceedings of the 19th ACM international conference on Multimedia, Scottsdale, Arizona, USA, 2011.