DOI QR코드

DOI QR Code

은닉 마르코프 모델을 이용한 동영상 기반 낙상 인식 알고리듬

Video Based Fall Detection Algorithm Using Hidden Markov Model

  • 김남호 (동양미래대학교 소프트웨어정보학과) ;
  • 유윤섭 (한경대학교 전기전자제어공학과)
  • 투고 : 2013.06.12
  • 발행 : 2013.08.15

초록

동영상에서 추출한 변수값을 은닉 마르코프 모델(Hidden Markov Model; HMM)에 적용한 새로운 낙상 인식 알고리듬을 제안한다. 개인간 낙상 양식의 차이나 유사 낙상을 실제 낙상과 구분하기 위한 기계 학습 방법으로 HMM알고리듬을 사용하였다. 비디오의 낙상 특징 변수를 얻기 위해 동영상의 광류를 구한 후 이를 주성분 분석 방식에 적용하여 움직임을 정량화하였다. 주성분 분석으로 얻어진 전체 움직임 벡터의 각도, 장단축의 비, 속도등의 조합으로 새로운 여러 종류의 낙상 특징 변수를 정의한 후 이를 HMM에 적용하여 결과를 비교, 분석하였다. 이들 변수들 중에 각도에 의해 얻어진 변수가 가장 좋은 결과를 보여 본 실험에서 91.5%의 민감도(성공 감지율)와 88.01% 의 특이도(실패 감지율)를 나타내었다.

A newly developed fall detection algorithm using the HMM (Hidden Markov Model) extracted from the video is introduced. To distinguish between the fall from personal difference fall pattern or the normal activities of daily living (ADL), HMM machine learning algorithm is used. For getting fall feature vector of video, the motion vector from the optical flow is applied to the PCA (Principal Component Analysis). The combination of the angle, ratio of long-short axis, velocity from results of PCA make the new fall feature parameters. These parameters were applied to the HMM and the results were compared and analyzed. Among the newly proposed various kinds of fall parameters, the angle of movement showed the best results. The results show that this parameter can distinguish various types of fall from ADLs with 91.5% sensitivity and 88.01% specificity.

키워드

참고문헌

  1. B. Kaluza, M. Lustrek, "Fall detection and activity recognition methods for the confidence project: a survey," in Proc. of the 12th International Multiconference Information Society 2008, vol. A, pp. 22-25, 2008.
  2. M. Popescu, Y. Li, M. Skubic, M. Rantz, "An Acoustic Fall Detector System that Uses Sound Height Information to Reduce the False Alarm Rate," in Proc. of 30th Int. IEEE EMBS Conf., pp. 4628-4631, Vancouver, BC, Aug. 20-24, 2008.
  3. A. K. Bourke, C. N. Scanaill, K. M. Culhane, J. V. O'Brien, and G. M. Lyons. "An optimum accelerometer configuration and simple algorithm for accurately detecting falls." in Pro. of the 24th IASTED international Conference on Biomedical Engineering, pp. 156-160, Innsbruck, Austria, Feb. 15-17, 2006.
  4. M. Kangas, A. Konttila, P. Lindgren, I. Winblad, T. Jamsa. "Comparison of low-complexity fall detection algorithms for body attached accelerometers." Gait & Posture, Vol. 28, issue 2, pp. 285-291, 2008. https://doi.org/10.1016/j.gaitpost.2008.01.003
  5. D. J. Willis. Ambulation Monitoring and Fall Detection System using Dynamic Belief Networks. PhD Thesis. School of Computer Science and Software Engineering, Monash University, 2000.
  6. B. Toreyin, Y. Dedeoglu, and A. Cetin. "Hmm based falling person detection using both audio and video", In Proceedings of IEEE International Workshop on Human-Computer Interaction, Beijing, China, 2005.
  7. H. Nait-Charif and S. McKenna, "Activity summarization and fall detection in a supportive home environment", In Proceedings of the 17th International Conference on Pattern Recognition(ICPR), vol. 4, pp.323-326, 2004.
  8. C. Rougier, J. Meunier, A. Arnaud, and J. Rousseau. "Fall detection from human shape and motion history using video surveillance", In Proceeding of IEEE Advanced Information Networking and Applications Workshops, vol. 2, pp. 875-880, 2007.
  9. A. Sixsmith and N. Johnson, "A smart sensor to detect the falls of the elderly", IEEE Pervasive Computing, vol. 3, no. 2, pp. 42-47, April-June 2004.
  10. Jianbo Shi and Carlo Tomasi, "Good Features To Track", In Proceedings of IEEE Conference on Computer Vision and Pattern Recognition(CVRP94), pp.593-600, June 1994.
  11. Gary Bradski and Adrian Kaehler, "Learning OpenCV: Computer Vision with the OpenCV Library", pp.329-332, O'REILLY, 2009.

피인용 문헌

  1. 단순 임계치와 은닉마르코프 모델을 혼합한 영상 기반 낙상 알고리즘 vol.18, pp.9, 2013, https://doi.org/10.6109/jkiice.2014.18.9.2101