The Journal of Korean Institute of Communications and Information Sciences (한국통신학회논문지)

The Korean Institute of Commucations and Information Sciences (한국통신학회)
권호 표지
DOI QR코드

DOI QR Code

A Compensation Algorithm for the Position of User Hands Based on Moving Mean-Shift for Gesture Recognition in HRI System

HRI 시스템에서 제스처 인식을 위한 Moving Mean-Shift 기반 사용자 손 위치 보정 알고리즘

  • Kim, Tae-Wan (Tongmyong University, Department of Computer Engineering) ;
  • Kwon, Soon-Ryang (Tongmyong University, Department of Electronic Engineering) ;
  • Lee, Dong Myung (Tongmyong University, Department of Computer Engineering)
  • Received : 2014.06.12
  • Accepted : 2015.05.14
  • Published : 2015.05.31
Download PDF
⟨ Previous Next ⟩

Abstract

A Compensation Algorithm for The Position of the User Hands based on the Moving Mean-Shift ($CAPUH_{MMS}$) in Human Robot Interface (HRI) System running the Kinect sensor is proposed in order to improve the performance of the gesture recognition is proposed in this paper. The average error improvement ratio of the trajectories ($AEIR_{TJ}$) in left-right movements of hands for the $CAPUH_{MMS}$ is compared with other compensation algorithms such as the Compensation Algorithm based on the Compensation Algorithm based on the Kalman Filter ($CA_{KF}$) and the Compensation Algorithm based on Least-Squares Method ($CA_{LSM}$) by the developed realtime performance simulator. As a result, the $AEIR_{TJ}$ in up-down movements of hands of the $CAPUH_{MMS}$ is measured as 19.35%, it is higher value compared with that of the $CA_{KF}$ and the $CA_{LSM}$ as 13.88% and 16.68%, respectively.

본 논문은 키넥트 센서 (Kinect sensor)를 탑재한 Human Robot Interface (HRI) 시스템에서 손 위치 데이터를 측정하여 제스처 인식 및 처리성능을 높이기 위하여 Moving Mean-Shift 기반 사용자 손 위치 보정 알고리즘($CAPUH_{MMS}$)을 제안하였다. 또한, $CAPUH_{MMS}$의 성능을 자체 개발한 실시간 성능 시뮬레이터로 이동궤적에 대한 평균 오차 성능개선 비율을 다른 보정 기법인 $CA_{KF}$ (Kalman-Filter 기반 보정 알고리즘) 및 $CA_{LSM}$ (Least-Squares Method 기반 보정 알고리즘)의 성능과 비교하였다. 실험결과, $CAPUH_{MMS}$의 이동궤적에 대한 평균 오차 성능개선 비율은 양손 상하 운동에서 평균 19.35%으로, 이는 $CA_{KF}$ 및 $CA_{LSM}$ 보다 각각 13.88%, 16.68% 더 높은 평균 오차 성능 개선 비율을, 그리고 양손 좌우 운동에서 평균 28.54%으로 $CA_{KF}$ 및 $CA_{LSM}$ 보다 각각 9.51%, 17.31% 더 높은 평균 오차 성능 개선 비율을 나타낸 것이다.

Keywords

References

  1. S. C. Kim and I. C. Park, "A study on the ubiquitous home network interface system by application of user's gesture recognition method," J. Korean of The Science of Emotion & Sensibility (KJSOS), vol 8. no 3, pp. 265-276, Sep. 2005.
  2. F. K. H. Quek, "Unencumbered gestural interaction," IEEE Multimedia, vol. 3, pp. 36-47, Dec. 1996. https://doi.org/10.1109/93.556459
  3. S. Mitra and T. Acharya, "Gesture recognition: A survey," IEEE Trans. Systems, Man, and Cybernetics, Part C: Applications and Reviews, vol. 37, no. 3, pp. 311-324, 2007. https://doi.org/10.1109/TSMCC.2007.893280
  4. G.R.S. Murthy and R.S. Jadon, "A review of vision based hand gestures recognition," J. Int'l Information Technol. and Knowledge Management (IJITKM), vol. 2, no. 2, pp. 405-410, July-December 2009.
  5. H. J. Kim, Kinect for windows 1: Implement of the innovative user interfaces of embedded devices (2012), Retrieved Feb. 1, 2015, from http://www.mdstec.com/bbs/press_room/upload/Kinect for Windows_4.pdf
  6. P. Doliotis, et al., "Comparing gesture recognition accuracy using color and depth information," in Proc. 4th Int. Conf. Pervasive Technol. Related to Assistive Environments (PETRA'11), no. 20, pp. 1-7, USA, May 2011.
  7. M. Yang, N. Ahuja, and M. Tabb, "Extraction of 2D motion trajectories and its application to hand gesture recognition," IEEE Trans. Pattern Anal. Machine Intell., vol. 24, no. 8, pp. 1061-1074, Aug. 2002. https://doi.org/10.1109/TPAMI.2002.1023803
  8. E. J. Holden, G. Lee, and R. Owens, "Australian sign language recognition," J. Machine Vision and Appl., vol. 16, no. 5, pp. 312-320, 2005. https://doi.org/10.1007/s00138-005-0003-1
  9. S. Y. Cho, et al., "Hand gesture recognition from kinect sensor data," J. Korea Broadcast Eng. (KJBE), vol. 17. no 3, pp. 447-458, May 2012. https://doi.org/10.5909/JBE.2012.17.3.447
  10. A. R. Kim and S. Y. Rhee, "Motion control of a mobile robot using natural hand gesture," J. The Korean Inst. Intell. Syst. (KJIIS), vol. 24, no 1, pp. 64-70, Feb. 2014. https://doi.org/10.5391/JKIIS.2014.24.1.064
  11. G. Welch and G. Bishop, An Introduction to the Kalman Filter (2006), http://www.cs.unc.edu/-welch/media/pdf/kalman_intro.pdf/
  12. T. W. Kim, "A study on the localization compensation algorithm using Mean kShift/ Kalman Filter in random walk/waypoint mobility model environment," Master's Thesis, The Graduate School of Tongmyong University, Feb. 2014.
  13. Konstantinos G. Derpanis, Mean shift clustering, Computer Vision Notes (2005), http://www.cse.yorku.ca/-kosta/CompVis_Notes/mean_shift.pdf/
  14. M. K. Jung and D. M. Lee, "Performance analysis of the localization compensation algorithm for moving objects using the least-squares method," J. KICS, vol. 39C, no. 1, pp. 9-16, Jan. 2014. https://doi.org/10.7840/kics.2014.39C.1.9