Author : JangYeol Lee 1
Date of Publication :18th July 2017
Abstract: The contact type gesture acquisition method is a method of specifying the threshold value and recognizing the gesture when the threshold value is exceeded. Since this method does not take into account other conditions depending on gender, age, and height, it is a problem of lowering the recognition rate when using the specified threshold value. In addition, the Threshold method can not guarantee the scalability of the Gesture because it has a limitation to add the operation of Gesture. In this paper, we apply the machine learning to the gesture to solve the problem of the existing threshold method and examine the recognition rate. We classify three gestures (rock-scissor-paper) obtained by contact device using multi class classification algorithm and conduct experiment to improve recognition rate. We perform three experiments to understand recognition rate improvement. First, we examine the recognition rate change according to the learning dataset size. Second, we examine the recognition rate when some sections of gesture data are learned and finally the recognition rate when smoothing is applied to gesture data. The results show that the recognition rate increases as the feature value of the gesture data becomes clearer, and the recognition rate of the gesture increases from 63% to 97%
Reference :
-
- Lu, Zhiyuan, et al. “A hand gesture recognition framework and wearable gesture-based interaction prototype for mobile devices,” IEEE transactions on human-machine systems 44.2 (2014): 293-299.
- Billot, Maxime, et al. “Age-related relative increases in electromyography activity and torque according to the maximal capacity during upright standing,” European journal of applied physiology 109.4 (2010): 669-680.
- Goldberg, David E., and John H. Holland. “Genetic algorithms and machine learning,” Machine learning 3.2 (1988): 95-99.
- Shvachko, Konstantin, et al. “The hadoop distributed file system,” Mass storage systems and technologies (MSST), 2010 IEEE 26th symposium on. IEEE, 2010.
- Zaharia, Matei, et al. “Resilient distributed datasets: A fault-tolerant abstraction for in-memory cluster computing,” Proceedings of the 9th USENIX conference on Networked Systems Design and Implementation. USENIX Association, 2012.