近年來，人機互動的重要性逐漸提升。手勢互動的行為逐漸佈滿我們生活之中，但也產生許多的問題和衍生多項的技術議題等待被解決。隨著科技迅速的演變與發展，人們正在使用各項技術來實現與機器溝通和互動的想法。

本論文使用單一視訊攝影機作為影像輸入的媒介，並建立即時運作的動態手勢辨識系統。首先，為了找出手部與頭部在畫面上所在的區域，我們對系統定義的使用者膚色做分割。接著透過幾何計算的方法取得手部和頭部的位置資訊讓系統與使用者進行即時的互動。最後，我們對取得的手勢資料作正規化處理，再使用機器學習中的隨機森林方法對動態手勢資料做訓練和辨識。

我們對五位不同的人蒐集六種NBA裁判手勢並建立資料庫，裡面共有600筆資料。實驗結果顯示對於這六種手勢的平均辨識率達98.5%。我們所提出的系統可達到即時辨識的表現，對於影格640×640大小的影像，整體運行表現平均為每秒30個影格。

In recent years, the importance of human-computer interaction has been gradually improved. Gesture interaction behavior has gradually filled our lives, but has also produced many problems and derivatives of a number of technical issues waiting to be resolved. With the rapid evolution and development of science and technology, human beings are using the technology to achieve communication and interaction with the machine.

In this paper, we used a single webcam as the medium for image inputs, and established a real-time dynamic gesture recognition system. First, in order to find out the area where the hand and the head are in the images, we segmented the user-defined skin color area set by the system. Then we used the method of geometry calculation to obtain information about the hands and head, providing information to allow the system to interact with the user instantly. Finally, we refined the gesture data by normalization, and then used Random Forest method in machine learning to deal with dynamic gesture data for training and recognizing.

We collected the six NBA referee gestures and created a database of five different people. There were 600 gesture data. The experimental results show that the average accuracy of the six gesture is 98.5%. The system we proposed can achieve the performance of real-time recognition, for 640 × 640 size images, and the overall average performance is 30 frames per second.

[1] H. H.S. Ip, K. C. Law, and B. Kwong, “Cyber composer: Hand gesture-driven intelligent music composition and generation,” Proceedings of the 11th International Multimedia Modelling Conference, Australia, pp. 46-52, Jan. 2005.
[2] “Cyber Glove Systems,” [Online] Available: http://www.cyberglovesystems.com/cyberglove-iii/ (accessed on June 20, 2017).
[3] D. Bowman, C. Wingrave, J. Campbell, and V. Ly, “Using pinch gloves™ for both natural and abstract interaction techniques in virtual environments,” [Online] Available: http://people.cs.vt.edu/~bowman/papers/pinch_gloves.pdf (accessed on June 20, 2017).
[4] H. Ishiyama and S. Kurabayashi, “Monochrome glove: A robust real-time hand gesture recognition method by using a fabric glove with design of structured markers,” Virtual Reality (VR), USA, pp. 187-188, Mar. 2016.
[5] N. Fujishima and T. Ietsuka, “Basic construction of a natural finger outline extraction system with a color glove,”15th International Conference on Computer and Information Science (ICIS), Japan, pp. 1-6: June 2016.
[6] O. De, P. Deb, S. Mukherjee, S. Nandy, T. Chakraborty, and S. Saha, “Computer vision based framework for digit recognition by hand gesture analysis,” IEEE 7th Annual on Information Technology, Electronics and Mobile Communication Conference (IEMCON), Canada, pp. 1-5, Oct. 2016.
[7] P. Barros, N. T. Maciel-Junior, B. J. Fernandes, B. L. Bezerra, and S. M. Fernandes, “A dynamic gesture recognition and prediction system using the convexity approach,” Computer Vision and Image Understanding, vol. 155, pp. 139-149, Feb. 2017.
[8] S. Ren, H. Wu, and Y. Wan, “Construction of virtual 3D life-like human hand for real-time human-computer interaction,” Seventh International Conference on Advanced Computational Intelligence (ICACI), China, pp. 345-350: Mar. 2015.
[9] G. Plouffe, A. M. Cretu, and P. Payeur, “Natural human-computer interaction using static and dynamic hand gestures,” 2015 IEEE International Symposium on Haptic, Audio and Visual Environments and Games (HAVE), pp. 1-6 Oct. 2015.
[10] “NBA台灣,” [Online] Available: https://nba.udn.com/nba/index (accessed on June 20, 2017).
[11] “NBA MINER,” [Online] Available: http://www.nbaminer.com/ (accessed on June 20, 2017).
[12] “HSL和HSV色彩空間,” [Online] Available: https://zh.wikipedia.org/wiki/HSL和HSV色彩空間 (accessed on June 20, 2017).
[13] H. Lahiani, M. Elleuch, and M. Kherallah, “Real time hand gesture recognition system for android devices,” 15th International Conference on Intelligent Systems Design and Applications (ISDA), Morocco, pp. 591-596, Dec. 2015.
[14] T. Pavlidis, “Contour filling in raster graphics,” presented at the Proceedings of the 8th annual conference on Computer graphics and interactive techniques, Dallas, Texas, USA, 1981.
[15] R. L. Graham and F. F. Yao, “Finding the convex hull of a simple polygon,” Journal of Algorithms, vol. 4, no. 4, pp. 324-331, Dec. 1983.
[16] L. Breiman, “Random forests,” Machine learning, vol. 45, no. 1, pp. 5-32, Oct. 2001.
[17] T. M. Oshiro, P. S. Perez, and J. A. Baranauskas, “How many trees in a random forest?,” International Workshop on Machine Learning and Data Mining in Pattern Recognition, Germany, pp. 154-168, July 2012.
[18] R. Genuer, J. M. Poggi, and C. Tuleau-Malot, “Variable selection using random forests,” Pattern Recognition Letters, vol. 31, no. 14, pp. 2225-2236, Oct. 2010.
[19] L. Breiman, “Bagging predictors,” Machine learning, vol. 24, no. 2, pp. 123-140, Aug. 1996.
[20] L. Breiman, “Out-of-bag estimation,” [Online] Available: https://www.stat.berkeley.edu/~breiman/OOBestimation.pdf (accessed on June 20, 2017).
[21] L. Breiman, J. Friedman, C. J. Stone, and R. A. Olshen, “classification and Regression Trees,” Chapman and Hall, London, United Kingdom, 1984.
[22] C. Cortes and V. Vapnik, "Support-Vector Networks," Machine Learning, vol. 20, no. 3, pp. 273-297, Sept. 1995.
[23] D. E. Rumelhart, G. E. Hinton, and R. J. Williams, “Learning Internal Representations by Error Propagation,” Parallel Distributed Processing: Explorations in the Microstructure of Cognition: Foundations, MIT Press, MA: Cambridge, vol. 1, pp. 318-362, 1986.